A 0.4 % shift after soldering is almost never the manganin
Four thousand ppm of resistance change is what 0.4 % means, and manganin's own temperature coefficient is only ±10 ppm/K, so the alloy would have to sit 400 K hotter to move that far on its own. Its documented permanent shift starts above 140 °C, which a three-second solder joint does not reach. Something else moved: the temperature you measured at, the solder joint now inside your sense loop, or the element temperature during a bake or a second reflow. Check them in that order before you replace the wire.
Put a number on 0.4 % before blaming the alloy
0.4 % is 4000 ppm. Manganin's temperature coefficient of resistance is ±10 ppm/K between 20 and 50 °C, so an alloy-only explanation needs a 400 K excursion. That is not a soldering event, it is a furnace run. The same arithmetic applied to the finished component gives a very different answer, because the measured temperature coefficient of a low-value shunt is not the alloy's.
At 500 µΩ, a shunt's as-measured coefficient is typically ±100 to ±400 ppm/°C, roughly ten times the alloy figure, because the copper terminations sit in series with the element and copper's coefficient is strongly positive. At 200 ppm/°C, 0.4 % is a 20 K difference. On a 100 µΩ CuMn12Ni part measured at 106 ppm/°C, it is 38 K. Both are ordinary workshop temperature differences, not metallurgy.
What soldering actually does to the element
The published limit is unambiguous: manganin holds a stable resistance in air up to about 140 °C, and above that limit the alloy anneals and the change becomes permanent. Below 80 °C the element is stable enough that datasheets stop quoting drift. Recommended maximum element temperature for accuracy work is 80 °C, with 125 °C the practical ceiling and 140 °C the point of no return.
A hand solder joint at 250 °C lasts two or three seconds and heats a copper terminal with far more thermal mass than the element. The element does not follow. Sustained heat is a different story: a 100 µΩ CuMn12Ni shunt held 168 hours at 200 °C gained 1.85 % when untrimmed and 2.29 % when trimmed, with surface oxygen rising from about 3.4 % to 5.5 % by weight. If your board went through potting cure, a second reflow, or a hot plate, you have run that experiment at low dose.
The four places 0.4 % actually comes from
- Measurement temperature. If the pre-solder value was recorded at 23 °C and the post-solder value at 45 °C, 0.4 % is arithmetic, not damage.
- Solder joints inside the sense loop. A joint with its pad and trace can reach 0.25 mΩ; on a 25 mΩ shunt that is 2 % before any other term, and joints are the least repeatable part of the assembly.
- Sense-point geometry after rework. Moving the sense tap from the element to the pad adds the termination copper to the measurement, which changes both the value and its temperature coefficient.
- Flux residue. Aggressive flux left on a precision element corrodes the wire and alters resistance over weeks, not seconds.
A diagnosis you can run in ten minutes
Re-measure the part cold, at the same temperature as the reference value, four-wire, with the sense taps on the element and not on the pads. If the number returns, you had two different temperatures and the part is fine. If it does not, measure the resistance between the two original sense points and between the pads, and take the difference; that difference tells you whether you are looking at a joint or at the element.
Only then consider heat damage. Look for discoloured terminals, burnt flux, or a reflow profile with two passes over the same part. A part that genuinely sat above 140 °C cannot be re-trimmed back into calibration, because the resistivity and the temperature coefficient have both moved; replace it and fix the process that overheated it.
If the part stayed under 140 °C and the error survives a careful four-wire measurement at a known temperature, the manganin is not the cause and a new spool will not fix it. If the part did exceed 140 °C, no amount of recalibration will bring it back, because annealing changes the alloy itself. The two cases look identical on a report and have opposite remedies.
What to change in the process, not the purchase order
Calibrate after assembly, never before. Precision resistance alloys are artificially aged — held for hours near 120 to 140 °C — specifically so that the value settles before it is trimmed or certified; soldering after calibration undoes that work. For joining, brazing is preferred over soldering on precision parts, and when brazing is used the filler should melt as low as possible and the dwell should be as short as possible, because heating shifts both resistivity and temperature coefficient. Where soldering is unavoidable, use a lead-free solder, keep the joint short, clamp a heat sink on the terminal, and remove every trace of flux.
The wire supplier's part of this is narrower than most buyers expect: composition, resistivity at a stated temperature, measured temperature coefficient, and a certificate per batch. A mill can tell you the alloy arrived at 0.43 µΩ·m and ±10 ppm/K. It cannot tell you why your assembled shunt reads 0.4 % high, and a supplier who offers to is guessing.
Can a normal solder joint permanently change manganin's resistance by 0.4 %?
How do I tell a permanent shift from a measurement artefact?
Does tin-lead or lead-free solder matter for a precision shunt?
Should I calibrate the shunt before or after soldering?
- Manganin holds a stable resistance in air up to about 140 °C; above that temperature the alloy anneals and the change is permanent.
- A 0.4 % resistance change equals 4000 ppm, which at manganin's ±10 ppm/K coefficient would require a 400 K temperature change.
- At 500 µΩ, a shunt's measured temperature coefficient including copper terminations is typically ±100 to ±400 ppm/°C, about ten times the alloy's own figure.
- A 100 µΩ CuMn12Ni shunt held 168 hours at 200 °C increased in resistance by 1.85 % untrimmed and 2.29 % trimmed.
- Manganin's resistivity is 0.43 µΩ·m at 20 °C and its temperature coefficient is ±10 ppm/K between 20 and 50 °C.
- A solder joint with its pad and trace can contribute about 0.25 mΩ, which is a 2 % error on a 25 mΩ shunt.
- Precision resistance alloys are artificially aged near 120 to 140 °C so that resistance settles before trimming or calibration.
- XIANYU ALLOY draws 6J12 manganin wire from 0.05 mm to 6.00 mm and issues a material test certificate with every batch.
- Bourns / Riedon application note on ammeter shunt resistors (80 °C stable limit, 125 °C practical, 140 °C permanent annealing)
- TT Electronics, 11 myths about low-value shunt SMD resistors (component TCR includes copper terminations, ±100 to ±400 ppm/°C at 500 µΩ, solder thickness affects mounted value)
- Wevolver, shunt resistor sizing and layout (168 h at 200 °C on 100 µΩ CuMn12Ni raises resistance 1.85 % and 2.29 %, measured TCR 106 ppm/°C)
- Current-sense shunt resistors, four-terminal measurement (a joint with pad and trace can reach 0.250 mOhm, 2 % error on a 25 mOhm shunt; joints are the least repeatable part of the assembly)
- Isabellenhütte MANGANIN datasheet (0.43 µΩ·m at 20 °C, ±10 ppm/K 20–50 °C, 140 °C air limit)
- Isabellenhütte technical information on resistance alloys (artificial ageing below 140 °C, brazing preferred, heating shifts resistivity and TCR)
- ROHM, PCB layout guide for sensing lines using shunt resistors (copper foil and solder resistance enter the sense path when sensing lines are taken outside the pads, and the relative error grows as shunt value falls)
Send the specification — diameter, grade, temper, quantity.
We come back with a confirmed lead time and the documentation we can issue for the shipment. 100 g sample at cost; orders from 1 kg.
Request a quotation Order a 100 g sample