So I take this subject of it being a 'real' 1.2th vs not being one right?
Folks, welcome to the world of binning. A standard practice throughout the solid-state electronic components industry. The only point that matters is:
do they reliably run at 1.2th +/- whatever %? Or for cpu's s: 1.6, 1.8, 3.2, 3.6, 3.8, whatever GHz. Not can you make them go faster. It's all about does in this case the chip package as presented performs reliably. BMch & AMT say yes. Works for me.
Intel, AMD, all GPU makers do it.Hell even power semi makers do it to some extent. With these and cpu, gpu's etc., from any given wafer there are some chips that can do very well (run fastest with lowest errors) and some not so good. They do work fine at slower speeds though so.... Test and separate by performance and sell them accordingly by their sustainable performance. Same chips, often from the same wafer. There is just enough confidence that the better/best will hold up to the higher speed so you sell them for more $$.
Gee, of the 4 Ant S1's I have rated 180GHs each, only 1 is over-clockable to 190GHs. So happens that it also still has the best HW rate doing it. Damn near but not quite zero. Push to 200 and errors skyrocket. As for the other 3 - @ out-of-the-box default of 180 they already have several x higher HW rate than the OC'd one so OC not even tried but gues what what - ya can eco-mode them as well. As in slow down. Guess what - the error rate drops substantially.
So, should I sue because Bitmaintech chose to label the Ants as 180 when they can run better at 160 or 140ghs (and as low as 0.6w/GH/s doing it vs 2w @180)? I think not. Goosed a bit from whatever you would call native mode(s) they run fine at 180. Some even faster.