When the spec and the drawings disagree, don't pick one and build the package around it. Write the conflict down the day you find it, send one RFI that cites both documents and proposes an answer, hold that item while the rest of the package moves, and put the RFI number on the transmittal so the reviewer sees what you built to.
Section 40 67 00, paragraph 2.03.A: "Enclosures located outdoors shall be NEMA 4X, Type 316 stainless steel." The panel schedule on E-601 lists CP-2 as NEMA 12, painted steel. The site plan puts CP-2 on the headworks platform, outside. The schedule is specific to the panel, the drawing set looked newer, and NEMA 12 was what you priced. So that's what you submitted.
Three weeks later: Revise and Resubmit. "Enclosure does not meet 40 67 00 2.03.A." Your shop has already cut the backpanel to fit the NEMA 12 box. Now you're re-quoting stainless, eating a lead time you didn't plan for, and asking the engineer to look at the same panel twice. One RFI on the day you read the section would have cost you a paragraph.
Most conflicts are visible on the first read
You don't find these at review because they're hidden. You find them at review because nobody read the spec section and the drawings at the same time. The engineer does exactly that, so plan on them finding it.
When you read a spec section for its submittal requirements, open the sheets that show the same equipment beside it: the P&IDs, the instrument schedule, the panel schedule and elevations, the one-line, and the network riser. Then check the places integrator conflicts usually sit:
- Enclosure rating and material. The section's rule for outdoor or corrosive locations against the panel schedule and the site plan.
- Instrument type and count. The section calls for a mag meter, and the instrument schedule says ultrasonic. The P&ID shows a level transmitter that isn't on the schedule at all.
- Power. The section asks for UPS-backed control power with a stated runtime, and the one-line shows no UPS circuit for that panel.
- Communications. The section names one protocol or media, and the riser shows another: EtherNet/IP against Modbus, fiber against copper.
- Quantities. One HMI per panel in Part 2, two on the elevation.
Every disagreement goes on a list the day you find it: the paragraph, the sheet and detail, what each one says in its own words, and the equipment it touches. Don't settle it in your head and move on. The one you settle in your head is the one that comes back stamped.
Check whether the book already decides
Before you write the RFI, look for an order of precedence. It usually lives in the supplementary conditions or Division 01. Some books say the spec governs over the drawings. Some say the reverse. Many say the documents are complementary and the more stringent requirement, or the greater quantity, governs. Most general conditions also require you to report a conflict you find before you do the affected work.
If the book plainly decides it, you may not need an RFI. Build to the governing requirement and say so on the transmittal: "CP-2 provided NEMA 4X 316 SS per 40 67 00 2.03.A, which governs over the E-601 panel schedule per 01 42 00 1.4." That puts the decision in writing for the reviewer.
If the book doesn't decide it, or deciding means picking the cheaper side, ask. "Which one is right" is the engineer's call, and an approval stamp doesn't make it for them. That line between proposing and interpreting is in submittal vs RFI.
What a tight RFI looks like
A busy engineer answers the RFI they can answer in one sitting. Give them one.
- One question. One conflict, one RFI. A bundle of five gets three answers.
- Both documents, cited. Section and paragraph on one side, sheet and detail on the other, each quoted in its own words.
- A proposed answer. "We propose NEMA 4X 316 SS for CP-2 per 2.03.A" turns their job into "confirmed."
- The impact and a date. What holds until they answer, and when you need it.
Compare these two:
Please clarify enclosure requirements for the control panels.
Section 40 67 00 2.03.A requires NEMA 4X, Type 316 stainless enclosures for panels located outdoors. Sheet E-601 panel schedule lists CP-2 as NEMA 12, painted steel. Site plan C-102 locates CP-2 on the headworks platform, outdoors. We propose providing CP-2 as NEMA 4X, Type 316 SS per 2.03.A. Panel fabrication and the CP-2 portion of submittal 40 67 00-002 hold pending this answer. Please respond by November 20.
The first one gets a reply that says "see spec." The second one gets "proposed approach accepted" in two days, because the engineer can check every claim in it without opening the drawings.
Attach the marked-up excerpt: the paragraph highlighted, the schedule row clouded. It saves them the search, and it shows you did yours.
Hold the item, not the whole package
An open RFI shouldn't stop everything. Build the rest of the package while you wait: the datasheets, the highlights, the drawings that don't touch the conflict.
What you do with the held item depends on the project:
- Partial submittals allowed. Send the package without the held item and say so on the transmittal: "CP-2 enclosure to follow, pending RFI-017." Check Section 01 33 00 first. Some projects reject incomplete packages outright.
- Partial submittals not allowed, or the conflict drives the whole package. Hold the package and chase the RFI. A package sent on a guess just turns the wait for an answer into a review cycle.
Either way, don't buy or fabricate the held item on a guess. And put the RFI's response date on your calendar. An RFI nobody chases is a guess with paperwork.
Show the reviewer the answer you built to
The answer comes back: "Provide NEMA 4X, Type 316 SS for CP-2 as proposed." Now make sure the reviewer of the submittal sees it, because they may not be the person who answered the RFI.
- Cite it on the transmittal. "CP-2 enclosure per RFI-017 response dated 11/12." One line, and the question "why does this differ from E-601" is answered before anyone asks it.
- Include the answered RFI. Put it behind the enclosure cut sheet or in a supplemental section, so the evidence is one flip away.
- Highlight the value the answer settled. The NEMA 4X rating and the 316 material on the cut sheet, so the reviewer checks it against the RFI in one glance. How to mark up cut sheets covers the markup.
- Price it if it changed the work. An answer that adds cost or time is a change, not a clarification. Raise it before you build, the same way you would a review comment that's really a change order.
When Rev 0 goes out with the RFI cited, the paragraph highlighted and the answer behind it, there's nothing left for the engineer to find.
How Submittal Kit handles this
Guessing feels faster than asking when asking means a Word template and an email thread. In Submittal Kit every project keeps an RFI Log beside its Submittal Log. You write the question, the spec and drawing references, and your proposed solution on one page, and attach the marked-up excerpt, which prints in the issued PDF. Mark sent assigns the next RFI number and freezes the PDF as the record of what went out. The log counts days from that date and flags the RFI overdue once it passes its response-expected date, so the held item chases itself.
When the answer comes back, drop the returned PDF and the answer is read onto the record, marked Changes the work, No change, or Needs more information. Then type the RFI number into the transmittal's Notes, which print in the compiled package, and upload the answered RFI to a Supplemental Documents or Custom Upload section so it rides in the package behind your cut sheets. See RFI basics and Configure sections.
Disclosure: I build Submittal Kit. The conflict list, the RFI and the transmittal note work with a spreadsheet and a PDF too.