When you order a sheathed thermocouple probe, one spec that quietly shapes how it performs is how the measuring junction relates to the metal sheath around it: grounded, ungrounded, or exposed. This choice trades how fast the probe responds against how electrically isolated and physically protected it is, and getting it wrong leads to either sluggish readings or noisy, ground-loop-plagued signals. This page lays out the three junction styles and the practical logic for choosing between them, particularly on points that feed a SCADA system.
Junction Types in one line: Thermocouple junction types describe how the measuring junction sits relative to the probe's protective sheath. A grounded junction is welded to the sheath for fast response but shares its electrical ground, an ungrounded (isolated) junction is electrically separated from the sheath for noise immunity at some cost in response speed, and an exposed junction sticks out beyond the sheath for the fastest response with the least protection.
In a grounded-junction probe, the thermocouple wires are welded directly to the tip of the metal sheath, so the junction is in intimate thermal contact with the sheath wall. Heat reaches it quickly through the metal, giving a fast response, and the tip is fully sealed and mechanically robust. The tradeoff is that the junction is electrically connected to the sheath, and therefore to whatever the sheath is bonded to - the process piping, a thermowell, or plant ground. That shared ground is the doorway through which electrical noise and ground loops enter the measurement.
An ungrounded, or isolated, junction is welded within the sheath but electrically insulated from it, floating inside the packed mineral insulation. Because a layer of insulation now separates the junction from the sheath wall, heat takes slightly longer to reach it, so the response is a bit slower than a grounded probe of the same size. In return, the junction is electrically isolated from the sheath and from process ground, which breaks the ground-loop path and dramatically improves immunity to electrical noise. Mechanically it retains the sealed, protected construction of a sheathed probe.
An exposed junction dispenses with the tip seal entirely: the welded junction protrudes beyond the end of the sheath directly into the process medium. With nothing between the junction and the fluid or gas, response is the fastest of the three by a wide margin. The cost is protection - the bare junction is vulnerable to corrosion, moisture, mechanical damage, and pressure, so exposed junctions are limited to clean, dry, low-pressure, non-corrosive environments such as measuring flowing air or clean gas temperature where speed is paramount.
The three styles line up neatly on a single axis. Exposed is fastest and least protected, ungrounded is slowest of the sheathed options but most isolated, and grounded sits in between with fast response and full sealing but shared grounding. There is no universally best junction; the right one falls out of what the application values more - the speed to catch a fast temperature change, or the isolation to keep the signal clean and free of ground loops.
Response time matters most where temperature moves quickly and control or safety depends on catching it. A fast excursion in a fired heater or a rapid transient in a gas stream is better tracked by a grounded or exposed junction, because a slow probe averages the change and reports it late. Where the temperature is slow-moving and stable, the response-time difference between the styles becomes largely academic, and other considerations take over.
Isolation matters most wherever the measurement can be corrupted by electrical noise or by a difference in ground potential between the process and the instrument. A grounded junction ties the thermocouple to process ground, and if the instrument end is grounded elsewhere, a ground loop can flow and inject error or even damage sensitive inputs. An ungrounded junction removes that path entirely. This is why, for signal quality on anything but the fastest measurements, the ungrounded junction is the conservative default.
For temperature points that feed an RTU or PLC and then a SCADA system, ungrounded junctions are usually the safer specification. Field installations sprawl across grounded piping, motor-driven equipment, and long cable runs, all of which create opportunities for ground loops and induced noise. An ungrounded, isolated junction keeps the thermocouple electrically clean so the signal the input module sees - and the value Merobix eventually historizes - reflects temperature rather than the electrical state of the plant. The modest hit to response time is rarely a problem for the slow thermal processes typical of oil and gas.
Reserve grounded junctions for the points where fast response genuinely earns its keep and the grounding can be managed, ideally with an isolated thermocouple input so the shared process ground does not close a loop through the instrument. Exposed junctions stay in their narrow niche: clean, dry, low-pressure gas or air measurements where speed dominates and there is no corrosive or high-pressure threat to the bare junction. Putting an exposed junction into a wet, dirty, or pressurized process is a short path to a failed sensor.
The junction type also colors how you read trouble from a remote dashboard. A temperature point that shows sudden noise, drift, or erratic jumps correlated with nearby equipment starting up is a classic ground-loop signature, and a grounded junction feeding a non-isolated input is a leading suspect. Knowing the junction style of each point turns that pattern in the historized trend into a concrete fix - move to an ungrounded junction or an isolated input - rather than an unexplained data-quality mystery.
Ungrounded is usually the safer choice for SCADA-connected points because it electrically isolates the junction from process ground and prevents the ground loops that inject noise into long field wiring. The tradeoff is a slightly slower response, which rarely matters for the slow thermal processes common in oil and gas. Reserve grounded junctions for points where fast response is essential and grounding is carefully managed.
An exposed junction protrudes beyond the sheath and contacts the process medium directly, with no sheath wall or insulation between the junction and the temperature it is measuring. A grounded junction still has to conduct heat through the sealed sheath tip. Removing that barrier is what makes the exposed junction respond fastest, at the cost of leaving the junction unprotected from corrosion, moisture, and pressure.
A grounded junction ties the thermocouple electrically to the sheath and to process ground. If the instrument end is also grounded at a different point with a slightly different potential, current can flow through the thermocouple circuit between those two grounds, injecting error into the reading. Using an ungrounded junction or an isolated thermocouple input breaks that loop and eliminates the problem.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.