Seems much of the eBay, Amazon and AliExpress banana adapters/plugs are bogus, and not copper. Not sure what the "gold" plating is on the ones we have, but doesn't take long to wear thru!! Not a fan of the "spindle" type plugs either, tend to give erratic contact, whereas the "bladed" types provide a more uniform contact.
Same here with my "gold plated" plugs but, wear issue aside, as long as the alloys used are the same for both prongs, it shouldn't matter as far as thermoelectric effects are concerned, especially if the sockets remain within a fraction of a degree of each other as appears to be the case with the SDM3065X (even better of course, when the sockets remain with a fraction of a degree of ambient in the case of battery powered hand held meters).
I don't recall seeing any "spindle" type plugs unless you mean the type used with cheap K type thermocouples, apropos of which I have two such which, aside from the error of the thermocouples themselves, seem to offer the minimum of socket to thermocouple wire temperature differential error. I can see just over a 1K drop over a couple of minutes after plugging these into the SDG3065X from stone cold (room ambient) as the heat of the socket spreads to the thermocouple wire connection deep within their red and black plastic shrouds.
Since the socket temperature as reported by the SDG3065X's isothermal temperature sensor is typically 3K higher than ambient (it's at the mercy of variations in air temperature thanks to the steady forced ventilation from its mercifully quiet cooling fan), that still leaves almost another 2K of error to contend with.
If you are getting good results with your voltage short compared to a copper wire voltage short that's good, then you know you have a good reference!! However, a voltage short and 4 wire resistance short are not the same, the 4 wire short must support a measurement current and the actual voltage sense positions are important, thus understanding the current path in the short and directing it to give the best result. Having a high thermal mass is both good and bad, good as less influenced by thermal environment perturbations, bad as takes longer to thermal stabilize.
Both the "Gold plated" banana plug shorting bar and the simple copper wire shorting link are equally good (or bad). The -4uV error seems to be a "Zero volt calibration error" which I can live with for the time being.
As far as K type thermocouple readings are concerned, that equates to a mere 0.1K error and I'm more interested in voltage levels in the 1 to 19 volt range where consistent and thermally stable readings are of greater importance rather than absolute accuracy. If I need to measure micro-volt values, it's not a great problem to subtract out the -4uV from my readings. Besides which, when you're so close to the zero volt end of even the 200mV scale, readings below 10uV are going to suffer uncertainties of several percent anyway.
Regarding the merits or otherwise of high thermal mass/inertia, I'm entirely with you on that one.
Trying to get a handle on "room temperature" is rather like holding onto a lively Eel with well greased up hands!
The best you can do is let the room cool down overnight and reach a reasonably stable equilibrium (and
don't turn any heating on - suffer for your hobby!)
Good temperature measurements are tricky, we have a bunch of K type thermocouples and a bunch of thermistors, and even a thermal imager to help. The thermal imager has another parameter (emissitivity) to consider for accurate measurements. Our bench DMMS are fixed and have limited physical range, the imager is portable, and we have a portable UT320D for quick differential measurements (handy for airflow effects).
Best,
I have an IR thermometer which surprisingly, gives readings within +0.1K of the cheap Atech in the attached images (it reads in increments of 0.2k). All bar the final image, as you can see from the time on the more expensive IT Works 'weather station' display, were recorded around 3pm, about four hours earlier than the final image.
The room temperature had dropped overnight from its rather toasty centrally heated 24 deg C value and the heating was left shut off for the day. This is the only condition where both the cheap and the expensive 'weather stations' will agree to within a fraction of a degree of each other (and also it would seem, the thermometer which normally hangs on the bay window wall where it usually reads two or three degrees lower).
The other inside/outside clock/thermometer is located on my workbench shelf, the other side of the room and the only thing of merit it can boast is having an outside sensor (currently blu-tacked to the alloy heat spreader plate it is sitting on, along with a bunch of K-type thermocouples) that tracks within 0.1K of the internal temperature readings.
This setup is the closest I have to a calibration reference, based on the "Wisdom of fools" consensus principle. Whilst it seems statistically improbable that they should all show the same error to within 0.2K of each other, this no guarantee that they aren't all off by the same 2 or 3 degrees worth of error. At best, they offer a sanity check as to the most likely truth of the matter until I can get my hands on a properly certified thermometer reference.
However, I have reason to think that the TM-902C is accurately showing its own ambient temperature (the yellow K-type plug is now a shorting link) to with 0.1K and therefore that of the room ambient, with just a mere 1mW of self heating to interfere with this ideal situation, since with my bestest K type thermocouple, it had read within 0.1K of the temperature of melting ice and 99 deg C with the thermocouple dipped into a kettle of boiling tap water at an absolute pressure of 1013mBar.
This is the second of my TM-902C purchases. The first one sported the red on/off push button and the rather annoying 10 or 15 minute auto-shut off feature of the later models. Whilst this one might seem to be a very early model with that slide switch (and no auto-shut off feature) it rather surprisingly, just like the later models, uses a pair of AAA cells in place of the expensive PP3 (9v block battery) and 3,3v LDO as used by the very early models.
Not only have I lucked out in getting one without the annoying and needless time out feature and a reliance on expensive PP3 batteries, I also seem to have lucked out on accuracy - the first one shows a 1.5 deg lower temperature with its input shorted out, so only worth hanging onto as an emergency spare.
Well, after some seven hours of suffering deprivation for the sake of this temperature experiment, I finally decided that not only had I'd seen enough, I'd also had more than enough
and turned the radiator back on, along with a 2.5KW heater to zoom the room temperature back to a more comfortable temperature before the CH shuts off for the night.
As a result, I'm now observing a thermometer reaction race. The favourite to win is of course, that cheap Atech now showing 22.3 deg, followed by the IT Works with a creditable 21.3, leaving the TM-902C in third place at 21.2 and the thermometer trailing in last place with a reading of just over 20 degrees.
The in and out readings on the other digital thermometer are now showing 23.1 and 20.8 respectively, just as I'd expected with such a rapid temperature rise. In the time it took to type that, the in temp is now neck and neck with the favourite at 23.5 which seems to be pulling ahead by 0.1 to 0.2 degrees. I could carry on this commentary but I think you've got the picture by now.
However, now that all the heating sources have been shut off, I can now report the final results:
The Atech came in first place at 23.9 followed surprisingly by the TM-902C (23.4) then the IT Works (23.1) with the thermometer close behind at 23 and the cheap clock in/out thermometer trailing last with in and out readings respectively of 22.4 and 21.7.
You might think this is all rather frivolous but it does highlight the problems of trying to get accurate room/lab temperature readings whilst it's being actively maintained to a "set temperature" by a heating system (or AC - not 'a thing' in most UK residences).