After reading through the 13 posts, I didn’t see one that covered my thoughts (though I saw some great suggestions for general info… ) So here’s my thoughts…
TC is basically just starting to come out of “the Wild West” stage, where it’s basically been a free for all, with various types of designs, implementations, and points of view, with respect towards how the science behind the application of the electronics theory is/should be applied (in the real world environment).
Given that, you also them have matters complicated by the (also real world) business aspects of trying to save money (from the perspective of available technology, as well as production environment), in which the two most common, and leading (but not the only) cases means cutting corners, and others raising prices.
In example, in the cases of most “off the shelf” Chinese brands (and this is only my opinion), they’re more likely to use currently available chipsets, and then tailor their use/feature set to the given application. Sometimes this means adding supporting chips/etc for secondary functions, etc. and sometimes, those may be omitted to save costs. Regardless, there’s so many ways to approach the desired goal (temp control), that you’re going to see different levels of both accuracy, and performance. Primarily due to the differences in implementation. Anyways…(and I’m really trying to keep it short, but I know it may not seem like it)
In simplest terms, additional battery drain in TC mode could theoretically be observed by the consumption of additional power consumed by the “resistance monitoring” circuit (which should include live, “on the fly” data calculations, as those numbers are crunched and then processed back through the chain to be accounted for by the power delivery section). But (IMO) this would be largely due to poor design, or component choices. As there’s really no other difference between TC and wattage mode.
Side note: part of the costs associated with the custom chips that YiHi and DNA have made also improve power efficiency. Which also explains why they are generally superior in not only the monitoring of the resistance changes, but in the yield with respect to lifespan of the batteries, compared to less complicated/efficient devices.