I need to connect a precision 100 Ohm resistor to my boards to calibrate the zero Centigrade point offsets for my PT100 circuit (the only other way being a water triple point cell AFAIK). So, I have selected a through hole mount precision resistor of 0.005% accuracy with a 2ppm temperature coefficient. This is specified at around 25 degC. The rated power is 600mW, but I will only be putting about 500uA through it at most.

The questions are:

Is this good enough? On paper it looks like it could calibrate to 13 milliKelvin equivalent accuracy.

How do I mount such a resistor? I intend to put it into a sealed box with terminal connections for the 4 wire PT100 and keep it in an enclosure that is temperature controlled to within a tenth of a degree. I am planning to have it hanging in the air with the two ends soldered to the terminals of the current carrying wires, and soldering connections for the sense wires close to the body. Any special precautions or advice on this? Has anyone done anything similar?

  • \$\begingroup\$ I think that it is good far enough good, even too much. \$\endgroup\$ – Marko Buršič Oct 5 '15 at 14:19
  • \$\begingroup\$ How are you going to measure the temperature. And beware of DC offsets..(unless you are going to measure those too.) Re: mounting you want the temperature sensor and PT100 to be well coupled thermally. \$\endgroup\$ – George Herold Oct 5 '15 at 16:08
  • \$\begingroup\$ The circuit is AC excited and all the common mode stuff nulled out. The actual PT100 is calibrated separately (for now). This is just to test electronics during production \$\endgroup\$ – Dirk Bruere Oct 5 '15 at 16:17
  • \$\begingroup\$ The box will be placed inside one of our company's temperature controlled incubators - I have one to use in the lab, so that's free. However, the 2ppm stability won't make much difference if I just have it on my desk, and the lab itself is air conditioned to within a degree or so \$\endgroup\$ – Dirk Bruere Oct 7 '15 at 9:10

I would hold out for a Z-series Vishay metal foil resistor, which are available in 100\$\Omega\$, nominal 0.2ppm/degree C.

Mount with the recommended (see mfr application data) lead length to the Y connections, heat sink during soldering and fill your box with dielectric foam (eg. Urethane) to keep air currents off. If you are paranoid, use low thermal EMF solder, but try to keep everything as symmetrical as possible in any case. You can/should get it calibrated by a lab if it's important.

  • 1
    \$\begingroup\$ Is crimp worse than solder? \$\endgroup\$ – Dirk Bruere Oct 15 '15 at 9:17
  • \$\begingroup\$ Interesting question. My initial reaction is to say, well of course it is worse, but upon reflection I'm not sure a good pair of crimps (at two positions along each lead) might not be slightly better. \$\endgroup\$ – Spehro Pefhany Oct 15 '15 at 12:01
  • 1
    \$\begingroup\$ I was wondering about oxide layers and limited contact points metal-to-metal compared to the area of contact of a blob of solder \$\endgroup\$ – Dirk Bruere Oct 15 '15 at 14:14
  • 2
    \$\begingroup\$ Resistance isn't very important since you'll be using a Kelvin connection. The exact point of contact makes a tiny difference because of the lead wire resistance but it's more important that it doesn't change. Thermal EMFs should be small and mostly cancel. I would use SbPb solder and clearly separate the force and sense taps on the leads (force connections further away from the resistor body of course)- and make sure they get to the outside connectors correctly (sounds obvious.. but I've seen mistakes made) \$\endgroup\$ – Spehro Pefhany Oct 16 '15 at 1:44
  • \$\begingroup\$ Spot welding might be an option to consider, I have no experience just thinking aloud. Obviously any connections run the risk of becoming thermoelectric if the metals are dissimilar and the junctions are not in close proximity. Keep a common slab of thermal conductor across all 4 joints if you can. Also remember to have the connections at the other ends of the sense wires thermally matched. \$\endgroup\$ – KalleMP Jan 12 '18 at 5:20

How do you mount it? Should be done without physical stress. This is both during any lead bending (even then check the component manufacturer's guidance on how close to bend to the body) and in service.

I specified bent loops for a Standards compliance machine design: through hole - loop -- body -- loop -- through hole Then assemble and check it works, finally throw it in the corner for 6 weeks to age. (in practice we found it made little if any difference whether kit was powered up, subtleties here too, some components age exponentially from last curie point)

  • \$\begingroup\$ TimC at the fringes of measuring accuracy weird stuff happens. Sometimes the magic that professional instrument companies discover is never published for competitive reasons. Dirk, for this reason I would try to reverse engineer some existing similar applications that are trusted to see if they have solved problems that you have not even thought of yet. \$\endgroup\$ – KalleMP Jan 12 '18 at 5:24

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.