Nous avons remplacé la jauge par un capteur à ultrasons haute température. Voici ce que je vous dirais avant que vous en achetiez un.

Pendant longtemps, our 85°C holding tank was measured the old-fashioned way: somebody walked up with a dip stick, cracked the hatch, and hoped the steam didn't fog everything before they could read the line. The sight glass was worse, half the time it was too filmed over to see through anyway. After a couple of close calls and one bad burn on a co-worker's forearm, the plant manager said enough. We needed level data we could read from the control room, and we needed it without putting anything inside that tank.

That's how I ended up researching high-temperature ultrasonic level sensors, which I'd honestly never given much thought to before. This is what the whole process taught me.

Why not just use a contact probe?

My first instinct was a float or a rod-type probe, because that's what I knew. But talk to anyone who runs contact instruments in hot media and you get the same story: Mise à l’échelle, Revêtement, corrosion, and the constant worry about thermal stress on the wetted parts. Our liquid leaves a film that bakes onto anything that sits in it. Cleaning meant pulling the probe, which meant downtime, which meant the production lead asking questions nobody wanted to answer.

Ultrasonic works differently. The sensor mounts above the surface and bounces a sound pulse off the material. Rien ne touche le liquide. Pas de flotteur à brouiller, no probe to pull for cleaning, no wetted part to replace. That one difference quietly solves a lot of problems before they start.

I'll be honest, bien que: radar was the alternative we looked at too. Non-contact radar handles foam and vapor much better, and if your process is full of steam you should probably skip ultrasonic and go straight there. But radar typically costs two to three times what a comparable ultrasonic unit runs, and for a fairly clean heated liquid it was overkill for us.

The models we actually priced

Once we settled on ultrasonic, the shortlist was pretty standard for the market:

• Siemens SITRANS Probe LU240. The integral unit with the high-temp transducer option is rated up to around 145°C, which covered our 85°C tank with margin. This was the one our integrator pushed from day one.
• VEGA VEGASON 63. Solide réputation, Configuration simple, but the standard version tops out around 80°C process temperature, so it only made sense for our cooler tanks.
• Endress+Hauser Prosonic. Nicely built and their local support is good, but quotes came in noticeably higher.

We priced the domestic units too, the unnamed Chinese brands that show up on every B2B site. They're genuinely cheap, I'm talking roughly ¥2,000 to 4,000 for a decent 6-8 meter range unit, and on a simple cool tank they work fine. But for a hot, vapor-heavy process we went with the Siemens, and the quote landed somewhere around ¥12,000 to 15,000 with the high-temp transducer and a remote display. Installation on top of that ran maybe another ¥1,500 to 2,000 for the bracket, Câble, conduit, and a few hours of our electrician's time. Total, including commissioning, came in under ¥18,000. The domestic unit would have been a third of that, but readings are only worth something if they're reliable, and I didn't want to learn that lesson the hard way at 85°C.

One thing to keep in mind: these are quote-priced instruments, and the spread between suppliers is wide. What you pay depends on range, temperature rating, Type de sortie, Enceinte, and whether you need remote electronics.

Installation was easier than I expected. The gotchas weren't where I expected.

Mechanically it's simple: bolt the sensor to the top of the tank, point it at the surface, run the 4-20mA wire back to the PLC, set your empty distance and span. Our electrician had it mounted in an afternoon.

The frustrating part was the physics. Ultrasonic needs a clean echo, and anything in the beam confuses it. We learned this the embarrassing way: our first attempt on a tank with a small mixer gave readings that jumped around like a bad altimeter, because the blades kept crossing the beam. Moving the sensor away from the mixer calmed it right down. We also had condensation dripping off the transducer face on cold mornings, which ate the signal until we angled the sensor slightly. If your surface has foam, expect trouble, foam eats the echo. If it's a bulk solids hopper with a steep material angle, the surface scatters the pulse and you have to pick the mounting spot carefully. None of this is in the brochure, but it's the difference between "Ça marche" et "it works here."

One more thing that surprised me: the dead zone. Every ultrasonic sensor has a blind spot right under the face, ours was about 0.3 Mètres, so you lose that top band of the tank. Fine for us, but if you need continuous reading right up to the brim, account for it in the range you buy.

Garantie, and how not to pay list price

Don't buy these at list price. That's not a negotiator flex, every vendor we talked to had room to move once we asked for a quote with the process details attached. Ask for a two-year warranty instead of the standard one-year, it was free or close to it everywhere we tried, and on a sensor bolted above a hot tank the warranty matters more than the discount. A couple of vendors threw in the mounting kit or a free site survey to close the deal. One even offered a trade-in on our old contact probe, which I didn't know was a thing until they mentioned it.

The real advice: send the vendor your actual conditions before you ask for pricing. Process temperature range. Vessel size and mounting opening. What you're measuring, et si la mousse, vapeur, poussière, or condensation is in the picture. The output you need and whether it's indoor or outdoor. Any enclosure or site standard you have to meet. The vendors who come back with a recommendation instead of a catalog page are the ones you can trust.

Where it actually earns its keep

We've now got these in four spots. The hot holding tank where the dip stick used to live. A hot washdown sump that's a pain to reach and where nobody wants a hand near the surface. A small hopper with dusty material that would have coated any contact device within a week. And a retrofit where the customer explicitly didn't want to open the vessel, so a non-contact unit was the only option that didn't mean a shutdown.

The day-to-day difference is hard to overstate. Instead of someone walking out to the tank every couple of hours, the operator watches the level trend on a screen. When it's dropping faster than expected, they know before it becomes a problem. That's the real value here: not the fancy spec sheet, but not having to stand next to a hot tank to find out what's in it.

What I'd tell you before you buy

If your process is a fairly clean hot liquid or a compatible solid, and you're tired of manual checks or contact probes that need constant cleaning, a high-temperature ultrasonic level sensor is a genuinely good tool. But buy it the way you'd buy a car you plan to keep: tell the vendor what you're actually measuring, get a quote on the full configuration, and confirm the temperature rating and vapor handling in writing.

And if a vendor says "sure, this one handles anything," Éloigne-toi. The right answer to "will it work on my foamy 100°C tank" est "let's check your conditions first," not a sales pitch. We learned that the slightly embarrassing way, and the sensor we ended up with has been reading steady for months.