Behaviour FG_03

FG_03 — the behaviour contract

Published from _docs/reference/fg-03-behaviour.md — edit it there, not here. Changes made in the wiki are overwritten on the next sync.

Slot 1 → slot 2 transfer with a 180° flip. Written from the shipped ST; the PLC is master for behaviour.

Source: Beckhoff/TwinCAT_1/PLC/PLC_1/Modules/13 FG_03/FG_03.TcPOU.

The reset and resume model is in gripper-station-reset.md §4 — safe positions, the snapshot rule, the handover tie-break. This page is what the group is and what its production cycle does.

---

Components

Nine symbols. Seven cylinders, all FB_DoubleSolenoidFeedback — two coils and two limits each — plus two grip sensors.

Arm	Axes	Gripper	Grip sensor
Arm 1 — the placing arm	Y_AxisX1, Y_AxisY1, Y_AxisR1	Y_Gripper1	B_Detect1
Arm 2 — the picking arm	Y_AxisX2, Y_AxisY2	Y_Gripper2	B_Detect2

The two arms name their axes the opposite way round

Every earlier version of this module had arm 1 wrong because of it.

Confirmed against the twin's Axis components, which is the only place it is written down:

Axis	Direction	Detail
Y_AxisX1	vertical	translates local Y, factor +0.001; max = raised
Y_AxisY1	horizontal	translates local Z, factor +0.001; max = transfer position, min = home over slot 2
Y_AxisX2	horizontal	translates local X, factor +0.001; max = transfer position, min = home over slot 1
Y_AxisY2	vertical	translates local Y, factor −0.001 — sign inverted, so a rising cylinder value descends; max = lowered

Two further traps in the same family:

• Arm 1 never lowers to place. X1 rises at step 60 and stays at max for the whole cycle. The 180° flip on R1 is what carries the part down into slot 2 — which is why no lowering step appears, and why looking for one makes the sequence read as though a step were missing.
• R1 retracted (0°) is home and the receiving orientation; extended (180°) is the placing orientation. It turns only at the transfer position: over slot 2 the sweep would go through whatever is in the slot. That is why step 140 traverses back before step 150 turns.

Production cycle

Full step table in gripper-station-reset.md §4. Three sub-sequences with a safe position between them:

Sub	Steps	Does	Ends at
1	20–40	pick out of slot 1	SAFE A — X2 min, Y2 min, arm 2 raised at home, holding
2	50–90	the transfer	SAFE B — X1 max, Y1 max, R1 min, arm 1 raised at the transfer position, holding, not yet rotated
3	100–170	flip, place, home	—

Exactly one moment has both arms on the same part. Step 80 waits for B_Detect1 before step 90 commands Y_Gripper2.Retract, so the part is never held by nobody. The twin's SequenceUnit3.cs separates those two motions with a 0.2 s dwell instead, which assumes the grip happened rather than confirming it.

Arm 2 goes home at step 100 before the rotate, not in parallel with it, so the rotating part can never meet arm 2 standing at the transfer position.

Why the MIL dwells are gone

SequenceUnit3.cs interleaves all fourteen motions with 0.2 s waits because the twin's animation has no limit switches to wait on. They are the animation's substitute for feedback, not a process requirement. This module waits on .Extended / .Retracted at every step, exactly as FG_01 waits on its gate limits. The one dwell kept is the 0.5 s settle at the head of the cycle, which has no sensor behind it.

Two rules that cost a part to learn

A cylinder limit is not a grip witness. Jaws closed on nothing reach .Extended sooner than jaws closed on a part, so .Extended is true either way. Only B_Detect1 / B_Detect2 say a part is held.

A de-energised double solenoid does not hold a gripped part in the twin. True of the real valve, false here — the Gripper releases the moment its cylinder moves at all. Arm 1 dropped a part mid-reset before this was understood. CyclicLogic now re-asserts the grip every scan in every state but Execute.

Both grip sensors are GripSensor.cs reading the OC Gripper's IsPicked — the gripper's own record of what it picked up, not an independent look into the jaws. They cannot report a part the gripper has silently lost. On real hardware these would be beams; the PLC code is written the same either way, but a diagnosis leaning on "the sensor would have seen it" is leaning on the optimistic witness.

Fault codes

ST_FG_03_Transfer_Status, published as HMI.Transfer.FaultCode. Table in gripper-station-reset.md §7.

Code	Meaning
1	Resetting could not separate the two grippers at the handover; nothing was moved
2	pick failed
3	handover failed
4	release failed inside Execute
5	started with a part held by an arm that is not at its safe position

HMI.Transfer.ResumePoint is published alongside and is not a fault — it is which sub-sequence the next Start would enter (1 full cycle, 2 handover, 3 place), derived every scan from the live sensors so the panel shows what pressing Start would do.

Safety

No Safety section is authored for this group in the twin. The handover interlock above is derived from the sequence itself, not from a safety statement, and is recorded as such. No other permissive is invented here. The cell chain (Machine.SafetyOk → AbortImmediate) applies.

Non-production states

Aborting and Stopping command no axis motion — a part is likely hanging in a gripper, and moving an arm is what tears it. They leave the axes alone; CyclicLogic keeps the grip asserted.

Resetting parks whatever is held at its safe position without opening the gripper, and the next Start resumes from there — step 0 derives the entry point from the two grip sensors. So an E-Stop mid-cycle costs a reset and a start, not a scrapped part and not a trip into the cell.

The pallet stays clamped by Index03 throughout, so the resumed cycle finishes the job and Index03 gets its OperationDone as usual — provided the operator is quicker than Index03's own operation timeout, which nothing in this group can extend.

Related

• gripper-station-reset.md — the reset model, safe positions, resume, the twin drop mechanism
• fg-05-behaviour.md — the same model with one gripper
• transport-behaviour.md — the Index↔group handshake