Smart Meter and Sensor Procurement Checklist: Itron, M18 Sensors, Profinet Encoders, and Thermal Cameras

Before You Read This

I manage procurement for a mid-sized water utility. Our metering and instrumentation budget is about $180,000 a year, and I have tracked every order for the past six years. This is the checklist I use before buying smart meters, sensors, or inspection tools. It covers the Itron water meters we have in service, an inductive sensor M18, absolute encoder Profinet options, and the Fluke vs FLIR thermal cameras debate that comes up in every maintenance shop.

If you work in procurement, operations, or engineering, this will save you some of the expensive lessons I learned the hard way.

The 8-Step Procurement Checklist

This checklist works for both large meter orders and small sensor replacements. Here is the short version:

1. Define the outcome before you define the product.
2. Calculate total cost of ownership, not the unit price.
3. Treat your inspection tools as part of the same budget.
4. Verify integration before you sign the PO.
5. Demand proof of accuracy, not promises.
6. Check the physical environment before finalizing the part number.
7. Stock spares for what you cannot afford to wait for.
8. Write down the reasons and the math.

1. Define the outcome before you define the product

Most buyers start with a product name, like 'I need an Itron water meter.' That makes sense because Itron has a strong reputation. But the product name is not a specification. Are you replacing old meters? Cutting billing complaints? Looking for remote shutoff? Each job points to a different configuration.

The same problem shows up with an inductive sensor M18. The M18 body size is standard, but output logic, sensing distance, cable length, and ingress rating can all be different. Ordering by diameter alone means you are betting that the spec is obvious. It is not.

I also think about how customers see the result. A meter that under-registers for a year does not just lose revenue. It makes customers question every bill. When I choose a meter, I ask how it will look from the customer side, not just how it looks in the warehouse.

2. Calculate total cost of ownership, not the unit price

Unit price is the wrong number to compare. Most buyers focus on the price per unit and completely miss programming fees, test certificates, and freight. In a recent quote comparison, Vendor A quoted $84 per Itron smart water meter. Vendor B offered a 'comparable' meter at $71. I almost went with B until I totaled the line items: programming fees, test certificates, extra freight, and a minimum-order surcharge. B was $9,400 more on a 500-meter order. That difference is hidden in fine print.

The same logic applies to small sensors. A $32 M18 inductive sensor becomes a $70 item after you add the right connector, mounting bracket, and spare cable. And if you compare an absolute encoder Profinet model to a generic encoder, remember that the generic one may require a protocol converter. The converter cost, plus the engineering time, can erase the price advantage.

3. Treat your inspection tools as part of the same budget

Thermal imaging is not a separate topic. I use a thermal camera to inspect switchgear, transformers, motor control centers, and enough loose connections to know how much damage a bad one can do.

The 'Fluke vs FLIR thermal cameras' question comes up every year. Both brands make reliable cameras. The difference is not sticker color. It is lens resolution, temperature range, NETD, and software costs. I chose a 320x240 model because the 160x120 version could not see a failing connection from the working distance we needed. That camera cost about $2,000 and caught a transformer problem that would have been a much bigger bill. Best spare line item I ever approved.

4. Verify integration before you sign the PO

Integration is where total cost goes to die. PROFINET is an established industrial Ethernet standard (IEC 61784-2), but that does not mean every device and PLC speaks the same version. I bought 24 absolute encoder Profinet units after checking voltage, shaft size, and price. What I did not check was the PLC firmware version. The older PLC needed a gateway. The gateway cost more than the encoders.

Ask the vendor to put compatibility in writing. If the salesperson says 'it should work' instead of 'it is supported on this PLC version,' that is an answer. Get a second one.

5. Demand proof of accuracy, not promises

For water meters, the smart part does not matter if the flow measurement is wrong. According to OIML R49 (oiml.org), certified water meters have defined maximum permissible errors. But the certificate only counts if the meter operates inside its rated flow range. I have seen meters installed at sites where the demand stayed far below the minimum flow rating. The meter looked fine on paper and billed wrong in practice.

Ask for the test certificate for the exact firmware version and model you are buying. That matters for any manufacturer, including an Itron smart water meter. Firmware changes happen, and a certificate for an older version is not proof for the new one.

6. Check the physical environment before finalizing the part number

An M18 inductive sensor is just a metal cylinder until you put it in a washdown area. An IP65 rating is not enough for a line that gets hosed down daily. You need IP67 or IP69K, or a sensor designed for chemical washdown. The extra $12 per sensor is nothing compared to a production line stopping because a sensor flooded. I learned that one the expensive way.

For an absolute encoder Profinet device, check connector orientation and shaft load. A great network protocol does not help if you cannot route the cable an inch from the motor housing.

This is also where the 'cheap option' story usually happens. The cheap part looks fine in the box. It fails in the first wet week. Then you pay twice for labor and order the good part anyway.

7. Stock spares for what you cannot afford to wait for

In Q2 2024, we switched vendors for a sensor line and cut unit cost by 14%. I still kept a stock of the old M18 sensors because retrofitting all the mounting brackets at once would have caused downtime. The spare inventory costs a few hundred dollars. An unplanned outage costs a lot more.

For water meters, I keep a small inventory of Itron meters for emergency replacements. The goal is not to hoard. It is to know the lead time and carry enough to cover a line break or a sudden cluster of failed meters.

What most people don't realize is that 'standard lead time' often includes buffer time vendors use to manage their production queue. It is not necessarily how long your order takes.

I still kick myself for not asking about that earlier. The 'standard lead time' on our first order was six weeks. The actual production time was closer to two weeks. By the time I understood that, I had already paid for expedited freight once.

So glad I double-checked the spare inventory before the last sensor order. I was one part number away from buying 40 sensors with the wrong output. The maintenance team would have loved that.

8. Write down the reasons and the math

After tracking six years of purchase orders, I found that about 18% of budget overruns came from expedited shipping and last-minute specification changes. We made a simple policy: no expedite without a written cost-benefit note from the person asking for it. Overruns dropped by about a third the next year. The fix was not a better product. It was better documentation.

Keep a spreadsheet with quote date, unit price, freight, integration costs, and certificate status. When somebody asks why you chose one option over another, you don't need a slide deck. You need a row and the math.

Looking back, I should have built that spreadsheet in my first year. At the time, I trusted the lowest quote and thought I was saving money. The rework and downtime ate the savings. If I could redo that decision, I would never compare only the price column again.

Final Thoughts

This checklist is not about being cheap. It is about spending where it matters. Customers see the meter on their property, the bill in their app, and the temperature of the water. That makes quality a brand issue, not just an engineering issue.

Prices change and product versions change. Verify current specs and ask for test certificates in writing. My reference quotes from the middle of 2024 are already stale, so yours will be too.