Build a Tank Level Instrument Acceptance Plan Before Installation

A tank level instrument can pass every line on the data sheet and still fail the day it goes into the vessel. The wrong mounting nozzle, an unexpected coating on the tank wall, or a signal nobody agreed to read a certain way will turn a clean sale into a service call during the next outage. A technical sales engineer who waits until installation to ask about process conditions, mounting, and who signs off on the reading is setting up rework, not a close. The fix is an acceptance plan built before the customer ever schedules the install, one that names the process data, the mounting facts, the signal use, and the people who will judge the result.

Why This Conversation Goes Wrong

Most of these conversations fail because the sales engineer moves too fast toward a model number. A process engineer says they need a level sensor for a tank, and the instinct is to match a part to the vessel size and call it done. That instinct skips the questions that actually determine whether the instrument works once it is welded into place. Resin, slurry, and other coating prone materials build up on probes and throw off readings on capacitance and some guided wave radar devices. Vessel geometry, nozzle placement, and internal obstructions change which technology even applies. Temperature and pressure ratings that look fine on paper can be wrong for the actual operating envelope, not the nameplate envelope.

The other failure is treating the sale as the finish line. A sales engineer who recommends a technology and moves on has left the customer without a way to judge, once the unit is mounted, whether it did the job. Nobody wrote down what a good reading looks like at high level and low level. Nobody named who checks it. When the plant finds a problem after the outage window closes, the sales engineer becomes the person explaining a failed installation instead of the person who built the criteria that caught it early.

How to Prepare

Before you get on a call about a specific tank, get the process data you need to ask sharp questions instead of generic ones. You need the material and its tendency to coat, foam, or build a vapor layer. You need the temperature and pressure range the vessel actually sees during normal operation and during any upset condition, not just the design rating. You need the vessel geometry, including nozzle size and location, agitators, and anything else that could sit in the beam path or probe path. You need to know what signal the plant wants out of the instrument, a continuous 4 to 20 milliamp output with HART, a simple high and low switch, or both.

You also need to know who signs off. On most plant projects that is not one person. A controls engineer usually owns the signal and how it feeds the control system. An operator lead usually owns whether the reading matches what people see and expect on the floor day to day. A process engineer usually owns whether the instrument fits the chemistry and the vessel. If you do not name these people before the install, you will not know whose approval actually closes out the acceptance plan.

This same habit, tying a decision to documented evidence and a named reviewer instead of a verbal description, shows up across manufacturing sales conversations. The post on how to Review a Tooling Deposit Request Against Drawing and Milestone Evidence covers a different scenario, releasing a payment only once a drawing and inspection data support it, but the underlying discipline is the same one you need here.

What to Say

Picture the process engineer, Drew, opening with a plan already half formed. Drew says, "We need a level sensor for this resin tank, and we want to install it during the next outage."

The sales engineer, Alina, does not reach for a model number yet. She says, "Before we get to a specific instrument, tell me what the resin does to a probe over time, and give me the temperature range and the vessel geometry it needs to handle." Drew explains that the resin tends to coat the tank wall and any internal probe surface after a few weeks of service.

Alina follows with the second half of the picture. "That changes which technology makes sense here. Now tell me how operators will use the signal day to day, and what reading the process team would call acceptable at high level and at low level." Drew names the controls engineer who owns the signal integration and the operator lead who will judge whether the reading matches what the floor sees.

Alina closes the conversation with the plan, not a part number. "Let's put those criteria on the drawing, along with the acceptance range at high and low level, and get our engineering team to review the application before we lock in a configuration." She has not certified anything from a five minute description, and she has not promised an installation date. She has given the customer a documented plan their own engineering process can use to judge the proposal once it comes back.

Mistakes to Avoid

Watch for a few habits that undo this kind of plan even when the intentions are good.

1. Naming a model number before the process data is confirmed. Once a customer hears a specific part, they anchor to it, and walking it back later costs you credibility.

2. Skipping the mounting and geometry questions because the tank sounds like others you have sold into before. Every vessel has its own nozzle placement, internal obstructions, and coating behavior, and assuming similarity is how the wrong technology gets specified.

3. Treating engineering review as your own team's job alone. The plan works because the customer's engineering process signs off on it, not because your team approved your own recommendation.

4. Leaving acceptance criteria as a verbal understanding instead of putting the range and the owners on the drawing. A verbal agreement about what counts as a good reading at high and low level will not survive a personnel change or a rushed outage schedule.

5. Letting the outage date drive the pace of the technical conversation. An outage window is real pressure, but installing an instrument that has not been matched to the process just moves the failure from the sales call to the startup.

How to Practice and Score It

A manager or the sales engineer working alone can rehearse this conversation before it happens with a real customer. Start from a vague opening, a customer naming a tank and an outage date with no other detail, and see whether the sales engineer's first move is a question about process conditions or a jump to a part number. Score whether they asked for material behavior, temperature and pressure range, and vessel geometry before naming a technology. Score whether they asked how the signal gets used and what counts as an acceptable reading at both extremes. Score whether they named specific roles, not a vague reference to "the plant," as the people who approve the plan. Score whether they proposed a drawing and an engineering review instead of a verbal handshake.

DealSpeak's manufacturing practice conversations let you run this exact scenario against an AI counterpart that plays the process engineer, pushes for a fast answer, and gives you a score on how well you held the line on evidence over assumption. You can find that scenario and others like it at Manufacturing practice conversationsalongside related situations such as the one covered in Define a Torque Tool Verification Plan Before Changing Assembly Controlswhere the same pattern of naming criteria and owners before a change goes into effect applies to a different piece of equipment.

Building the acceptance plan before installation costs you one extra conversation up front. It saves you a return trip, a missed outage window, and a customer who no longer trusts your recommendation on the next tank.