I/O Lists, Instrument Indexes, and Loop Drawings: The Other Half of the Controls Submittal

A panel drawing package shows how the panel is built. It does not show what the panel is connected to, and on a process job the reviewer cares about that at least as much. The instrument index says what devices exist. The I/O list says which signals the controller sees and where each one lands. The loop drawings say how every wire gets from the device to the controller. Three documents, one set of tag numbers, and the review of a process control submittal is largely a check that the three agree with each other and with the P&IDs.

Most of the pain here comes from treating the three as separate deliverables produced by separate people at separate times. They are one database seen three ways.

The tag is the key

Everything hangs off the tag number, and the tag grammar is standardized. ANSI/ISA-5.1 defines instrumentation symbols and identification: a functional identification made of letters (the first letter is the measured variable, succeeding letters the function) and a loop number. LIT-401 is level, indicating, transmitter, loop 401. LSH-401 is the level switch high in the same loop. LIC-401 the controller. Devices in one loop share the loop number, and that shared number is what lets a reviewer trace a signal across all three documents without a lookup.

The tags come from the engineer's P&IDs. You do not invent them, rename them, or "clean them up." A spec I hold says it directly: all components, including field-mounted and rear-of-panel instruments, shall be tagged with the item number and nomenclature as shown and the instrument index in the contract documents. A tag on your I/O list that does not appear on the P&ID is the first thing a reviewer will circle.

The instrument index

One row per tagged device on the job, field and panel. Sometimes called the instrument list. There is no ISA standard defining its columns; ISA's own guidance describes it as an alphanumeric listing of all tag-marked components, each referencing the drawings and documents for that device. It originates with the engineer during design and often arrives as part of the contract documents, and the integrator maintains it from there.

The columns real specs ask for, drawn from several process control sections:

That last column is where the index and the I/O list meet. The data sheet reference is the other tie: ISA-20 specification forms, one per device, are what the spec means when it says "provide data sheets in accordance with ISA 20 and annotated for features proposed for use." The process control section I hold asks for exactly that, and for a similar format where ISA-20 has no form for the device. ISA has since retired the 1981 standard in favor of its technical-report form set, but the specs still say "ISA 20" and reviewers still want that sheet.

The I/O list

One row per signal the controller handles, including spares. It is a subset of the index, filtered to the tags that produce or receive a signal, plus the columns that only make sense on the controller side. Real specs are specific about those columns. One county's process control section wants, for every active and spare point: the full ISA tag, the type of I/O, the physical location as panel name, PLC or remote unit number, rack, slot, and point, the I/O address, the point name, and the terminal strip and number. Another owner's tag map adds signal source, PLC address, and HMI address.

The columns that matter and why:

• I/O type: the four classes below, plus network or soft points where the spec tracks them.
• Signal: the electrical form, because a 4 to 20 mA loop powered from the panel is wired differently from one powered by a field device, and a reviewer checks.
• Address: rack, slot, channel. This is the column that proves the hardware in the panel package can carry the points on the list, spares included.
• Terminals and wire numbers: the connection to the loop drawings and the panel wiring diagrams.
• Range and units, alarm setpoints, fail state: what the program will do with the point; these tie the list to the control narrative.
• Drawing reference: the loop drawing sheet for the point.

Spares are rows, not a footnote. Specs state spare I/O as a percentage per type, commonly ten to twenty percent installed and wired to terminals, sometimes with a minimum count, plus spare rack space, and one owner I have read rejects unwired spares outright. The reviewer counts. An I/O list with eleven analog inputs on a sixteen-point card and no spare rows is an I/O list that fails the spare check even if the card has room.

One of my specs also draws the line on ownership: the input/output list in the contract documents identifies the required points and is to be used for coordinating signals between equipment under this and other contracts, but may not be totally inclusive, and verification is the supplier's responsibility. Read that as the spec's I/O list being a starting point you own from award onward, and any point a vendor's package needs that isn't on it being yours to add at no cost.

The loop drawings

One sheet per loop, showing every device, wire, terminal, and power source between the field instrument and the controller. The standard is ANSI/ISA-5.4 Instrument Loop Diagrams, from 1991 and still the one specs cite; the process control section I hold references it by name, and so does every one I have read. Its minimum content, paraphrased: loop and component identification per ISA-5.1; a word description of the loop's function; interrelation with other loops; every point-to-point interconnection with the identifying numbers or colors of cables and conductors, including junction boxes, terminals, and grounding; the general location of each device (field, panel, rack, termination cabinet, I/O cabinet); energy sources with voltage and circuit or disconnect numbers; the process connection; and actions or fail-safe positions. The standard makes signal ranges, I/O addresses, and manufacturer and model optional, and most specs make some or all of them mandatory in the same sentence that cites the standard.

The generic loop drawing, one typical sheet with "similar for all loops," is the classic rejection. One public spec puts it in one sentence: submittal of generic diagrams will not be acceptable. Each loop gets its own sheet with its own tags, terminals, and wire numbers, because the loop drawing is also the loop check form: the field technician walks it wire by wire at startup and signs it, and a field testing section I hold requires the loop checks to be documented on the loop drawings themselves and submitted, with test type, date, problems, and signatures.

What the spec demands, and when

The process control section's submittal article usually lists all three by name, and its sequencing rules are stricter than the panel package's. Patterns from the specs I hold and have read:

• The instrument index and data sheets come with the field instrument product data, early, because instruments are long-lead.
• The I/O list comes with the panel shop drawings, because the reviewer checks the hardware against it.
• Loop drawings come later, once the panel design has settled, and one spec sets the deadline as at least 90 days before the first planned startup.
• The program documentation comes after the panel drawings are approved, and factory testing may not start until the programming submittal is approved.
• As-built versions of all three go in the O&M manual and the record documents.

One test plan submittal I hold will not allow any Division 13 equipment to ship until it is approved, and the factory test in another spec is a one hundred percent I/O point test, spares included, against the previously submitted I/O list. The list you submit is the list you will be tested against.

The mismatches that get them bounced

The reviewer is checking agreement, so disagreement is the finding. From owner standards that publish their rejection criteria, and from every returned package I have seen:

• Tags that don't match the P&ID. Renamed, re-numbered, or missing. Fatal.
• Missing terminal numbers, wire tags, junction box hops, or conduit IDs on loop drawings. The sheet has to be walkable.
• Missing range, polarity, or alarm thresholds, so the loop drawing cannot be checked against the index.
• No spare rows, or spare percentage below the spec's number, or spares present on the list but not wired in the panel.
• Missing surge protection or intrinsic-safety barriers on loops that leave the building or enter a classified area, when the spec requires them at every field loop including spares.
• Loop drawings that are really electrical schematics: panel-centric, no field side, no process connection.
• The control narrative referencing tags that aren't on the I/O list, or the I/O list carrying points the narrative never uses.

Every one of those is a consistency failure between documents that should have been generated from one source. That is the practical lesson: keep one table of tags with every column above, and produce the index, the I/O list, and the loop drawing title blocks from it. Edit the table, regenerate the documents, and the three cannot disagree. The panel drawings that sit beside these three are covered in what a complete panel drawing package includes, and the whole package in what goes in a control system submittal package.

One example

Three rows of one loop, as they appear in each document:

Instrument index
LIT-401  Wet well level      P&ID I-3  Loop 401  Radar transmitter
         Vendor model RLT-80  0 to 20 ft  4-20 mA  Field, LS-4
         Data sheet DS-401   I/O: PLC-4 R0 S3 CH2

I/O list
LIT-401  Wet well level   Analog input   4-20 mA, loop powered from PLC-4
         PLC-4 rack 0 slot 3 ch 2   TB3-14/15   0-20 ft   HH 18.0 ft
         Fail: last value, alarm    Loop dwg L-401

Loop drawing L-401 (one sheet)
  LIT-401 field, cable C-401 pair 1 (wire 401+ / 401-) to JB-4 terminals 7/8,
  cable C-4A pair 3 to LS-4 panel TB3-14/15, SPD-3 between, to analog input card
  slot 3 ch 2. Power: 24 VDC from PS-1, fuse F3-2. Range 0-20 ft.
  Fail-safe: n/a (input). Loop check: ______ date ______ by ______

Same tag, same loop number, same terminals, same address, in all three. A reviewer with the P&ID open finds LIT-401 on sheet I-3, checks the index for the model and range, checks the I/O list for the card and the spare count on that card, and walks the loop drawing to the terminal. Five minutes, no comments.

Disclosure: I build Submittal Kit, whose Equipment Register holds each panel's tagged components and feeds the submittal and the O&M manual from one list; it does not generate an I/O list or loop drawings today, and the three documents above stay in your own table. Keeping that table as the single source is the practice; the tool is whatever regenerates from it.