Short answer. Cold junction compensation corrects a thermocouple reading for the temperature where the thermocouple wires meet the instrument's copper terminals. Without it, a Type K at 100 °C with terminals at 25 °C reads 75.9 °C, not 75 °C. The error is close to the terminal temperature, not equal to it, and a wiring fault can push it high as well as low. Cold junction compensation (CJC) corrects a thermocouple reading for the temperature of the point where the thermocouple wires meet the instrument's copper terminals. A thermocouple produces a voltage set by the temperature difference between its tip and that point, so the instrument measures the terminal temperature, converts it to the equivalent thermocouple voltage, and adds that voltage before converting the total to temperature. Without cold junction compensation, a Type K thermocouple at 100 °C with its terminals at 25 °C reads 75.9 °C, not 75 °C (in Fahrenheit, 212 °F with 77 °F terminals reads about 169 °F). A Type T at 300 °C with the same 25 °C terminals reads 282.8 °C, only 17.2 K low. The error is close to the terminal temperature, but not equal to it, and wiring faults can push it high as well as low. Reviewed by the Focusens Engineering Team, 24 September 2026. How cold junction compensation works on a K-type thermocouple Thermocouple reference tables give the voltage a thermocouple produces with its reference junction at 0 °C. The National Institute of Standards and Technology (NIST) ITS-90 thermocouple database publishes those tables and the reference functions behind them, and IEC 60584-1 standardises the same functions. A field instrument never has a 0 °C reference, so it measures the real one and corrects for it. Worked through for a Type K thermocouple with its tip at 400 °C and the instrument terminals at 30 °C, using the NIST Type K table: The thermocouple delivers the difference between 400 °C and 30 °C: 16.397 mV − 1.203 mV = 15.194 mV. The instrument's terminal sensor reads 30 °C. The table voltage for 30 °C is 1.203 mV. The instrument adds the two: 15.194 + 1.203 = 16.397 mV. It looks up 16.397 mV in the Type K table: 400 °C. Skip step 3 and 15.194 mV looks up as 371.5 °C, which is 28.5 K low rather than 30 K low. The correction is made in millivolts, not in degrees. Beckhoff's thermocouple basics (EPP3314-0002 documentation) states that the compensation "is carried out with the voltages and not with the temperature". Adding 30 °C to 371.5 °C would give 401.5 °C, a 1.5 K error created by the arithmetic alone. Without compensation, the error is close to the terminal temperature, not equal to it A thermocouple's sensitivity changes with temperature. Near room temperature Type K produces about 40.5 µV per kelvin; at 400 °C about 42.2 µV/K; at 1000 °C about 39.0 µV/K. Type T rises from about 40.7 µV/K at 25 °C to about 58.1 µV/K at 300 °C. The voltage lost by ignoring the terminals is fixed by the terminal temperature, but the number of degrees it represents depends on th...

Short answer. A 10k NTC thermistor reads 10,000 Ω at 25 °C (77 °F), whichever curve it follows. Away from 25 °C the curves separate. At 0 °C a Type 2 reads 32,651 Ω, a Type 3 29,490 Ω, a 10k 3950 part about 32,755 Ω and a 10k 3435 part about 27,700 Ω. A 10k NTC thermistor reads 10,000 Ω at 25 °C (77 °F), whichever curve it follows. Away from 25 °C the curves separate. At 0 °C a Type 2 thermistor reads 32,651 Ω, a Type 3 reads 29,490 Ω, a 10k 3950 part reads about 32,755 Ω and a 10k 3435 part about 27,700 Ω. The chart below gives all four curves from −40 to 125 °C, followed by the error a controller shows when it expects one curve and gets another. Reviewed by the Focusens Engineering Team, 24 September 2026. 10k thermistor resistance chart, −40 to 125 °C °C °F Type 2 (10K3A1), Ω Type 3 (10K4A1), Ω 10k 3950, Ω 10k 3435, Ω −40 −40 336,098 239,828 343,633 204,700 −30 −22 176,803 135,233 179,267 118,500 −20 −4 97,006 78,930 97,840 71,020 −10 14 55,304 47,549 55,595 43,670 0 32 32,651 29,490 32,755 27,700 10 50 19,903 18,787 19,938 18,070 20 68 12,493 12,268 12,501 12,110 25 77 10,000 10,000 10,000 10,000 30 86 8,056 8,196 8,051 8,301 40 104 5,325 5,594 5,315 5,811 50 122 3,601 3,893 3,588 4,147 60 140 2,487 2,760 2,473 3,011 70 158 1,752 1,990 1,738 2,224 80 176 1,256 1,458 1,243 1,668 90 194 916 1,084 904 1,267 100 212 678 817 668 975 110 230 510 624 500 760 120 248 388 482 380 598 125 257 341 426 333 533 Sources, one per column. Type 2: Measurement Specialties (now TE Connectivity) 10K3A1 datasheet. Type 3: TE Connectivity 10K4A1 datasheet. Both columns agree with Vishay Curve 1 and Curve 9 in the Vishay NTC resistance/temperature conversion tables to within 0.3%, under 0.1 K, from −20 to 100 °C. 10k 3950: typical values from the Zhuhai Hongjiacheng HNTC-103F3950FB datasheet, a B25/50 = 3950 K material, rounded to the ohm. 10k 3435: Semitec 103JT datasheet, B25/85 = 3435 K, rounded to the ohm. Fahrenheit values are converted and rounded to the whole degree. The 3950 and 3435 columns each come from one manufacturer's material. Other materials sold under the same beta number follow a slightly different curve, because a beta value fixes two points and the rest of the curve depends on the ceramic. Plotted from the datasheet values in the table above. The lower panel is the same four curves measured against Type 2. Ask for the part's own resistance table; the beta alone does not define it. Which curve is which: Type 2, Type 3, 3950 and 3435 "Type 2" and "Type 3" are building-automation names, not a standard. In many North American HVAC output tables, a table labelled 10K-2 or Type II carries the values of the 10K3A1 curve (Vishay Curve 1), and a table labelled 10K-3 or Type III carries the 10K4A1 curve (Vishay Curve 9). The beta values below put all four materials on the same footing. Common name Element curve Published beta B25/50 B25/85 B0/50 10K Type 2 (10K-2, Type II) 10K3A1, Vishay Curve 1 B25/85 = 3976 K 3936 K 3977 K 3892 K 10K Type 3 (10K-3, Typ...

Short answer. A KTY84-130 and a PT1000 both sit in a motor winding, both are two-wire, and both land in the same one to three kilohm band. They cross near 208 °C and disagree everywhere below it. Fit the wrong one and a winding at a real 150 °C can display as 87 °C, or a cold motor can display as 100 °C and refuse to start. The consequence is not symmetrical, and that asymmetry is the point of this page: A KTY84 element read by a PT1000 input under-reads. At a real winding temperature of 100 °C it reports 0 °C. At a real 150 °C it reports 87 °C. The overtemperature alarm never fires. A PT1000 element read by a KTY84 input over-reads. At a real 0 °C it reports 100 °C. The machine trips on a cold motor and will not start. One failure mode is loud and stops production. The other is silent and removes the thermal protection while every instrument on the panel shows a plausible number. Why this is happening now Siemens has migrated its motors and spindles from the KTY84-130 to a PT1000. The migration document is public, and its wording about the two characteristics is unambiguous: the KTY sensor "is discontinued and no longer available in the market", the PT1000 is the successor to be used in its motors and spindles, and "the temperature characteristic is not compatible!!!". See Change of temperature sensor KTY84-130 on PT1000. The same document sets out the transition: delivery of the new configuration from 2016, phase-out from around 2018, product discontinuation in 2020, product cancellation in 2028, with ten years of spare-part availability. Motors and spindles without a DRIVE-CLiQ interface received new article numbers; those with the interface kept theirs because the sensor type is handled on the far side of the interface. Affected lines named in the document include SIMOTICS S, M, L and T, the 2SP1 spindles, SINAMICS S and G, SINUMERIK and SIMOTION. What the document does not do is say how far wrong the reading goes when the sensor and the input disagree. That is a number an engineer needs when a rebuilt motor comes back with the wrong element in it, or when a spare from the shelf predates the changeover. The rest of this page is that number. The two curves, side by side Temperature KTY84/130 typical (Ω) PT1000 per IEC 60751 (Ω) −40 °C 359 842.7 −20 °C 424 921.6 0 °C 498 1000.0 25 °C 603 1097.3 50 °C 722 1194.0 100 °C 1000 1385.1 150 °C 1334 1573.3 200 °C 1722 1758.6 250 °C 2166 1941.0 300 °C 2624 2120.5 KTY84/130 typical values from Table 7 of the NXP KTY84 series datasheet. PT1000 values computed from the reference function in IEC 60751, with A = 3.9083 × 10⁻³, B = −5.775 × 10⁻⁷ and C = −4.183 × 10⁻¹² below 0 °C; the same function produces our PT1000 resistance table. KTY84 values from the NXP data sheet, PT1000 computed from IEC 60751. Through the range a motor actually operates in, the KTY84 always reads lower. Read down the two columns and the shape of the problem appears. The PT1000 starts high and climbs gently at about 3.85 Ω per kelvin...

Short answer. On a 2-wire PT100, 1 Ω of loop resistance reads as 2.56 °C. On a PT1000 the same ohm is 0.26 °C, because the element is ten times more sensitive. Twenty-five metres of AWG 24 copper is about 4.2 Ω of loop, which is 10.8 °C on a PT100 and 1.1 °C on a PT1000. The conversion is one division. What is usually missing is not the ratio but the rest of the arithmetic: which gauge, over what length, and which part of the error a third wire or a commissioning trim actually removes. This page gives the number, the table, and the two cases where the usual rule of thumb is wrong. The one division you need A platinum RTD changes resistance at a rate set by IEC 60751. At 0 °C that rate is 0.3908 Ω per kelvin for a PT100 and 3.9083 Ω per kelvin for a PT1000. Divide the loop resistance by that slope and you have the error in degrees. Element dR/dT at 0 °C (Ω/K) at 25 °C at 100 °C 1 Ω of loop reads as PT100 0.3908 0.3879 0.3793 2.56 °C PT1000 3.9083 3.8794 3.7928 0.256 °C Slopes calculated from the Callendar-Van Dusen coefficients in IEC 60751. The slope falls slightly as the sensor heats, so the error a given cable causes grows a little with process temperature. Full values are in the PT100 and PT1000 resistance tables. The slope is why the factor of ten exists and why it never goes away. It is not a tolerance, a grade, or a manufacturing difference. A PT1000 simply moves ten times further for the same degree. What your cable actually adds On a 2-wire connection both conductors are in series with the element, so the loop is twice the one-way length. Copper resistance at 20 °C from the standard AWG table gives the numbers below. Every solid line is a PT100. The whole PT1000 range fits inside the shaded band at the bottom. Cable, one way Loop Ω, AWG 18 / 22 / 24 PT100 error °C PT1000 error °C 5 m 0.21 / 0.53 / 0.84 0.5 / 1.4 / 2.2 0.05 / 0.14 / 0.22 10 m 0.42 / 1.06 / 1.68 1.1 / 2.7 / 4.3 0.11 / 0.27 / 0.43 25 m 1.05 / 2.65 / 4.21 2.7 / 6.8 / 10.8 0.27 / 0.68 / 1.08 50 m 2.10 / 5.30 / 8.42 5.4 / 13.6 / 21.5 0.54 / 1.36 / 2.15 100 m 4.19 / 10.59 / 16.84 10.7 / 27.1 / 43.1 1.07 / 2.71 / 4.31 Two-wire connection, copper at 20 °C, both conductors counted. Read the three numbers in each cell as AWG 18, AWG 22 and AWG 24. In metric terms those are roughly 0.82, 0.33 and 0.20 mm². A 10 °C error on a Class A PT100 is roughly seventy times the element's own tolerance at 0 °C. At that point the sensor grade you paid for has stopped mattering. Why 3-wire cancels less than people assume A 3-wire connection does not remove lead resistance. It measures one conductor and subtracts it from the other two, which works only while all three conductors are identical. They usually are, to within a fraction of a percent, when the cable is a single homogeneous run. The mismatch shows up somewhere else: a crimp that is not quite the same as its neighbour, a junction box where one conductor was extended, a terminal that has oxidised on one leg only. What survives on a 3-wire r...

Short answer. A healthy KTY84/130 reads 577-629 Ω at 25 °C and 970-1030 Ω at 100 °C. Anything near 1000 Ω at room temperature is a different part, not a faulty one. The reading that actually separates a good sensor from a counterfeit is taken above 200 °C, where it must still be climbing toward 1641-1803 Ω. The most-read question about this sensor on NXP's own community is some version of "how do I tell whether it still works". The published answer points at Table 7 of the data sheet and warns not to exceed 10 mA. That is correct, and it is not enough. A table of nominal values does not tell you which readings mean a dead part, which mean the wrong part, and which mean the wiring. This page fills in that gap. What a healthy reading looks like NXP publishes an acceptance window, not just a nominal curve. Table 7 of the data sheet gives minimum, typical and maximum resistance at every step from -40 to 300 °C. It lives in a PDF, which is why most people quote the typical column and stop there. Here is the window itself, for the KTY84/130. Temperature (°C) Min (Ω) Typ (Ω) Max (Ω) TC (%/K) -40 340 359 379 0.84 0 474 498 522 0.79 25 577 603 629 0.74 100 970 1000 1030 0.61 150 1282 1334 1385 0.54 200 1641 1722 1803 0.48 250 2046 2166 2286 0.44 300 2456 2624 2791 0.29 KTY84/130 at a 2 mA sense current, from Table 7 of the NXP KTY84 series data sheet. Bold rows are the two measurements that decide most questions. The shaded band is what a good part is allowed to read. Note how much it widens above 150 °C. The last column is the one to notice. The temperature coefficient falls steadily with temperature, and then drops sharply between 250 and 300 °C, from 0.44 to 0.29 %/K. The curve flattens at the very top of the range. That has a practical consequence. Near 25 °C the part moves about 4.6 Ω per degree, so a 10 Ω reading error is roughly 2 °C. At 300 °C the acceptance window itself is 335 Ω wide, which is about 44 °C. A KTY84 used at its ceiling is not a precision sensor, whatever its grade. One reading at room temperature tells you which family it is Before judging whether a part is healthy, settle what it is. The families do not share a reference point, so a single ambient reading separates them. Reading at 25 °C What the part is about 600 Ω KTY84 series about 1000 Ω KTY81 or KTY83, 1 kΩ series about 2000 Ω KTY81 2 kΩ series, or a KTY10-5 / -6 / -62 / -7 about 1100 Ω a PT1000, not a KTY at all open circuit, or a few ohms failed part, or you are measuring the cable A PT1000 reads 1097 Ω at 25 °C by the IEC 60751 reference function, close enough to 1000 Ω that a quick look confuses it with a 1 kΩ KTY. Check at a second temperature before deciding. The last two rows matter more than they look. Fitting a PT1000 where the drive expects a KTY84 makes a cold motor report about 100 °C and refuse to start; the reverse is worse, and what a drive reads with the wrong sensor works through both directions in kelvin. Why your meter will not read the data-sheet number e...

Short answer. A KTY84 cross-reference table tells you which part has a similar curve. It does not tell you what will break. Four things do: the KTY84 family is specified at R100 = 1000 Ω and not at 25 °C, its ceiling is 300 °C and no silicon successor reaches it, the KTY83 and KTY84 parts are polarised while KTY81 and KTY82 are not, and any change of base resistance means the controller has to be relinearised. The KTY silicon sensors are being withdrawn, and the designs that used them are now being re-sourced under a deadline. Siemens states in its own product note that the KTY sensor is discontinued and no longer available in the market, and TI publishes an end-of-life replacement note for the KTY81 and KTY82. Most of what is published about this is a table of "equivalent" part numbers. That table is useful, and we publish one ourselves in the KTY83-110 cross-reference guide. It is also the part of the job least likely to cause a field failure. The four problems below are the ones that come back after the new parts are already fitted. Break 1: KTY84 is specified at 100 °C, not at 25 °C Nearly every other thermistor on your bench is quoted at 25 °C. KTY81 and KTY83 follow that habit: R25 = 1000 Ω or 2000 Ω. KTY84 does not. Its 1000 Ω figure is the value at 100 °C. At 25 °C a KTY84/130 sits at about 603 Ω. If an engineer reads "1 kΩ" on a KTY84 line and pairs it with a 1 kΩ part specified at 25 °C, the two curves are roughly 40 % apart at room temperature. The bridge still reads something. It reads the wrong thing. These are the published values for a KTY84/130 at a 2 mA sense current, with the tolerance NXP states in kelvin at each point. Note how the tolerance widens at the top of the range. Temperature (°C) Min (Ω) Typ (Ω) Max (Ω) Tolerance (K) -40 340 359 379 ±6.48 0 474 498 522 ±6.07 25 577 603 629 ±5.84 50 694 722 750 ±5.59 100 970 1000 1030 ±4.90 150 1282 1334 1385 ±7.10 200 1641 1722 1803 ±9.71 250 2046 2166 2286 ±12.73 300 2456 2624 2791 ±22.12 KTY84/130 at a 2 mA sense current, from Table 7 of the NXP KTY84 series data sheet. The 1000 Ω nominal appears at 100 °C, not 25 °C. The practical check takes one minute: put the candidate part and the original on the same bench at room temperature and compare readings before anything is soldered. A 40 % gap is visible immediately. Worth reading the last column honestly. At 300 °C the tolerance is ±22 K and the acceptance window is 335 Ω wide. A KTY84 reaches that temperature, but it is not a precision sensor when it gets there, and a design that needs accuracy at the ceiling was never well served by this part. Break 2: the 300 °C ceiling that silicon does not follow KTY84 runs to 300 °C. That is why it ended up in motor windings, spindles and brake-adjacent sensing in the first place. It is also the specification that no drop-in successor matches. Four replacement routes and what each actually covers. The shaded band is the range a KTY84 design may be using and no silicon PTC can reach. TI names t...

A PTC thermistor is a resistor whose resistance rises as it gets hotter. PTC stands for positive temperature coefficient. That single definition covers two devices that behave nothing alike and are bought for opposite reasons. A silicon PTC changes by about 0.76% per kelvin across its whole range, smoothly, and you use it to read a temperature. A ceramic switching PTC sits under 100 Ω all the way up to its rated point, then crosses 1330 Ω within five degrees and 4 kΩ within fifteen, and you use it to trip something. Specifying one when you needed the other is the most common mistake in this category. Published 10 September 2026. Reviewed by the Focusens technical department. The two families side by side Both are sold as "PTC thermistors". These are real figures from two Focusens series, one from each family. Silicon PTC (FTY series) Ceramic switching PTC (MZ6 series) What it is for measuring temperature detecting that a limit was crossed Curve virtually linear flat, then a step at the rated point Coefficient 0.76 %/K at 25 °C not meaningful; specified as thresholds instead Range −40 to +150 °C rated points from 60 to 190 °C Resistance at 25 °C 1000 Ω nominal for the 1 kΩ types ≤100 Ω single, ≤300 Ω triple Resistance above the rated point rises smoothly, R100/R25 = 1.67 typical ≥1330 Ω at +5 K, ≥4 kΩ at +15 K Rated current at 25 °C 5 mA, 10 mA maximum continuous driven by the relay, ≤30 V DC working Response 60 s thermal time constant ≤5 s Output you get a resistance you convert to degrees a state, below or above the limit Specifications from the Focusens product catalogue. The 60 s figure for the silicon series is quoted against different media on the series page and the element table, so confirm the medium against a part code before designing a control loop around it. A response time without its medium is not a specification. Read the resistance rows together. A silicon PTC moves from 1000 Ω to about 1670 Ω across 75 degrees. A switching PTC moves from under 100 Ω to over 4000 Ω across fifteen. The first is a measurement. The second is a switch that happens to be made of ceramic. Built from the Focusens catalogue threshold figures. The switching curve carries no usable temperature information below its rated point. Which one you actually need Answer three questions in order. Do you need a number or a decision? If a controller has to display, log, or regulate against the temperature, you need a sensor, and that means a silicon PTC, an NTC or an RTD. If something has to happen when a limit is crossed and nothing has to happen before that, a switching PTC does it with no conversion and no calibration. Where does the device sit? A sensor sits on a signal input. A switching PTC for protection usually sits inside the thing being protected, embedded in a motor winding or clamped to a transformer core, wired to a relay that only reads two states. Does it have to reset itself? A switching PTC returns to its low resistance once it cools, which is why it ...

Three PTC thermistors go into a three-phase motor, one per phase, embedded in the winding heads and wired in series to a protection relay. Below the rated temperature the string sits at a few hundred ohms. When any one of the three passes its rated point, the string crosses about 4 kΩ and the relay drops the contactor. The numbers in that paragraph are the whole design. This page gives them, explains why a series string triples every threshold, and covers how to choose the rated temperature against the motor's thermal class. Published 10 September 2026. Reviewed by the Focusens technical department. Why a PTC and not a current-based relay A thermal overload relay infers winding temperature from line current. That inference holds while the motor is cooled normally and fails in the cases that actually burn windings: blocked ventilation, high ambient, a clogged filter, rapid start-stop duty, or a motor running at rated current in a hot room. Current looks normal. The winding does not. An embedded PTC measures the winding itself, so none of those cases hide from it. It does not replace a current-based device, because it cannot see a locked rotor fast enough and it does not protect against short circuits. Standard practice is to run both. A PT100 in the winding gives a continuous temperature reading instead of a threshold, which is what you want on a large machine with a monitoring system. The PTC is the cheaper last line: no scaling, no calibration, two states. What the relay actually reads A switching PTC is specified by resistance at fixed offsets from its rated temperature, written TK. The relay is built around those offsets rather than around a temperature. MZ6 series values for rated points between 90 and 160 °C, TK tolerance ±5 °C: Measured at Single element Three in series 25 °C ≤100 Ω ≤300 Ω TK − 5 K ≤550 Ω ≤1650 Ω TK + 5 K ≥1330 Ω ≥3990 Ω TK + 15 K ≥4 kΩ ≥12 kΩ From the Focusens MZ6 series datasheet. Rated points of 60 to 85 °C and 165 to 190 °C use a different threshold set with a wider tolerance band, so read the row for the TK you are ordering. Every value in the right column is exactly three times the left. That is the property the whole scheme depends on: a relay whose trip point was set for a single element still reads a three-element string correctly, because the string scales linearly and only one element has to go high to carry the sum past the threshold. Threshold values from the Focusens MZ6 datasheet. Only two wires leave the motor regardless of how many elements are in the string. Why cold resistance tells you almost nothing The 25 °C row is the row people test against, and it is the least informative one on the page. Cold resistance is specified only as a maximum, ≤100 Ω for a single element and ≤300 Ω for a string. A healthy element and a marginal one both sit well under it. Nothing about the value at 25 °C predicts whether the step will land where it should, because the step is a property of the ceramic's transition and the c...