Skip to content

MAGNET_COMMISSIONING.md

Latest commit

 

History

History
175 lines (137 loc) · 8.94 KB

MAGNET_COMMISSIONING.md

File metadata and controls

175 lines (137 loc) · 8.94 KB
Raw

Ledex / PVA 1.6 commissioning handoff

Status 5/11/2026: Checkpoints 1–4 done on the bench 11/10 (wiring verified, 1 A fuse fitted, coil current vs AO0 measured: 10 V gives 0.31 A at minimum Imax; table in calibration/README.md). Mechanical mount done 1/11: adjustable holder, link collar and link bar printed and installed with the 16 mm window constrainer; the plunger is linked to the joystick shaft (the steel-washer target is rejected, so the washer wording in the checkpoints below is historical). Ledex-off hand check passed (full return to home, no binding). Checkpoint 5 is partial: the powered direction and release check and a gram-scale force measurement at the cube are still to do. Plunger must stay fitted and seated (without it the coil gives no usable pull). Model estimate for the link (simulations/ledex_femm/, README "Plunger link"): resistance at the cube from home to a 15 mm push of about 2.3 → 0.6 gf at 1 V, 13 → 4 gf at 5 V and 40 → 11 gf at 10 V; measure before use. LabVIEW magnet safety (see the end of this page) is still required before any nonzero mag.

Status 27/09/2026 (historical): the PVA 1.6 0–10 V driver arrived at the training rig; driver and coil were not yet wired together.

This page is the handoff for the rig-PC AI agent. Guide the user through one checkpoint at a time, record actual measurements, and update docs/PROJECT_PLAN.md after every completed step. Do not skip ahead after a failed measurement.

Safety and Sunday scope

  • Keep the Class 3B laser keyed off throughout.
  • No animal and no experimental parameter file during commissioning.
  • Secure the driver and Ledex before power. Keep fingers, loose steel tools and the washer target away during first energisation.
  • Never open or alter the mains side of the power adapter.
  • The PVA can source up to 3 A even though the Ledex is nominally only about 0.58 A. Install a 1 A inline fuse in the 12 V positive lead before powered testing. If no fuse is available, stop after unpowered inspection and wiring preparation.
  • The printed Ledex holder is not proven to provide the heatsinking assumed by the continuous-duty coil rating. Use only short pulses with long cool-downs today.
  • Stop for a blown fuse, unstable current, failure to release at 0 V, rapid heating, smoke/odour, unexpected LED state, binding or incorrect force direction.

Primary driver datasheet: https://www.hydraulikshop.ch/ftp/PVA_1.6_0919.pdf

The shop description contains a 0–10 mA typo. The ordered product and datasheet variant are 0–10 V differential input. Confirm the physical unit says 0–10 V; stop if it is the 4–20 mA variant.

Known hardware limits

  • Ledex 195224-230: 12 VDC, 7 W, approximately 20.7 ohm, nominal current approximately 12 / 20.7 = 0.58 A.
  • NI PCIe-6321 AO0: ±10 V range but only ±5 mA drive. Power-on/off can glitch as high as 2 V for 500 ms, so keep the PVA Enable open until AO0 is initialised and measured at 0 V.
  • AO0 command: Dev1/ao0, SCB-68A terminal 22.
  • AO0 reference: AO GND, SCB-68A terminal 55.
  • The coil must be powered by the reserved 12 V supply through the PVA, never by AO0.

PVA terminal map

  • D: +12–36 V supply (+Ub)
  • F: power ground
  • E: earth / chassis function
  • J: Enable; below 1 V is disabled, above 4 V is enabled
  • K: internal +10 V reference; unused here
  • M: signal ground for the local-reference circuit; unused here
  • H: positive 0–10 V command input
  • G: negative command input
  • C: positive coil output
  • A: negative coil output

The LED should be red while disabled and green while enabled. Do not guess or jumper F, G and M together. Do not improvise a mains-earth connection to E.

Checkpoint 1 — unpowered preflight

With all power disconnected:

  1. Confirm the driver is the 0–10 V model.
  2. Measure and record the Ledex resistance. Expected cold value is approximately 20.7 ohm; roughly 18–22 ohm is a reasonable initial acceptance band.
  3. Check that neither coil lead is shorted to the Ledex case.
  4. Identify the adapter's positive and negative low-voltage conductors with the DMM.
  5. Measure PVA H–G resistance as an initial AO-input screen. Stop if below 2 kohm, because 10 V / 2 kohm already reaches the NI AO0 5 mA limit. A high or unstable semiconductor reading is only a preliminary screen.
  6. Secure the PVA and Ledex mechanically.

Record the values before proceeding.

Checkpoint 2 — wiring, still unpowered

Wire:

12 V PSU +  ── 1 A inline fuse ──> PVA D
12 V PSU −                       ──> PVA F

SCB-68A terminal 22 / AO0        ──> PVA H
SCB-68A terminal 55 / AO GND     ──> PVA G

PVA C                            ──> Ledex lead 1
PVA A                            ──> Ledex lead 2

PVA D ── removable switch/jumper ──> PVA J Enable
                                      (leave open)

Leave K and M unused. Handle E only as part of an established lab chassis/functional-earth scheme. Do not add a random external flyback diode across C/A: the PVA is itself a PWM inductive-load driver, and an external diode may alter current regulation and release time.

Checkpoint 3 — disabled power and AO command

  1. Disconnect the coil for this checkpoint and leave Enable J open.
  2. Set Imin and gain / Imax to their minimum positions without forcing the trimmer end stops. Leave dither at the factory position initially.
  3. Power the 12 V supply. Measure about 12 V across D–F; the PVA LED should be red. Stop if it is green with J open, or if there is heating or smell.
  4. In NI-MAX, initialise MagnetPush_Dev1 to 0.0 V. Before connecting H, measure terminal 22 relative to 55 and confirm 0 V.
  5. Connect H/G, then verify that commands of 0 V and small positive steps are reproduced across H–G. Never command a negative voltage.

If the DMM can safely measure the command-input current, measure it at 10 V while the output remains disabled; it must be below 5 mA. Recheck the DMM lead sockets before returning to voltage mode.

Checkpoint 4 — establish coil current with short pulses

  1. Power off and insert the DMM in series with the coil on its suitable high-current range.
  2. Keep the steel target away. Reconnect the coil, initialise AO0 to 0 V, power the PVA, then close the removable Enable jumper.
  3. At a 0 V command, coil current must be effectively zero. Adjust Imin toward zero if needed. Stop if it energises strongly at 0 V.
  4. Apply approximately 0.5-second pulses at 0.5, 1, 2, 3, 4, 5 V, with at least 10 seconds off between pulses. Record command voltage, current, supply voltage, sound, release and temperature trend.
  5. Continue toward 9–10 V only if current is stable, the coil remains cool and current stays below 0.58 A. Measured current is authoritative. At the published minimum gain, rough expectations are 0.30 A at 5 V and 0.54 A at 9 V, but the 12 V supply leaves little driver headroom.
  6. Establish and record the maximum allowed AO command that does not exceed 0.58 A cold.
  7. Confirm that returning AO0 to 0 V removes current promptly. Open Enable before changing wiring.

Checkpoint 5 — mechanical direction

Mount the Ledex on the home side of the object so the plunger pull (through the link to the shaft) opposes the forward push. Start with the lowest proven command. Verify:

  • resistance is backward, not assistive;
  • the object, collar and link cannot collide with the Ledex or the constrainer over the full push;
  • the spring returns the object;
  • the 1D mechanism does not bind;
  • current returns to zero and force releases at 0 V.

Sunday's successful endpoint is a repeatable, low-duty pull at known current with clean release. Do not claim force calibration without a force gauge or load cell. A later calibration must map AO command to force at the actual operating gap and characterize heating across repeated trials.

Existing LabVIEW behavior and later code work

No LabVIEW edit is required for initial electrical commissioning:

  • NI-MAX task MagnetPush_Dev1 owns only Dev1/ao0.
  • The main VI currently wakes the magnet loop at trial Case 2, writes scalar parameter mag, waits magtime, then writes 0 V.
  • Safe startup already writes MagnetPush_Dev1 = 0 V.
  • Normal Stop has been validated to end all loops without DAQ errors.
  • Every current training/bench parameter file keeps mag = 0; leave them that way.

Before animal use or any nonzero experiment file, the canonical VI still needs:

  1. a clamp from 0 V to the measured safe maximum;
  2. guaranteed 0 V on success, timeout, Stop and DAQ/error paths;
  3. logged command, calibration version, onset and offset;
  4. verification of magnet-loop trial indexing;
  5. later replacement of the Case-2 timed pulse with movement-onset activation held until trial end, matching the axial-resistance design.

Record actual wiring, measurements and outcomes in RIG_INVENTORY.md and calibration/README.md, and update docs/PROJECT_PLAN.md after each completed checkpoint.