I may yet have to eat my own words regarding the explosive yield.
Most of the photos we've seen of the piles of crumbled shipping containers are not showing the devastation precisely at ground zero. They are slightly off the edge of the crater by about 50m.
I have now seen new aerial photos of the site, in particular one looking straight down from directly above the crater ... which I estimate to be roughly 75m(!) in diameter. As mentioned the explosion pretty much cleared a further 50m beyond the rim of the crater, and only then do the piles of containers start. I have now also seen photos of what appear to be large fields of 'container confetti', which has been thrown pretty far from the explosion. That would then be the shredded containers from close to ground zero that I asked for earlier in the thread.
The diameter of the crater expected by detonating a stationary sphere of solid TNT of mass W kg at ground level above soft, sandy soil is roughly:
D(m) = 0.8 * W1/3
So if the 21000kg yield of TNT equivalent is accurate, we'd expect the crater to be about 22m in diameter. Conversely, creating a crater of 75m in diameter would require 420t (420'000kg) of TNT equivalent.
However these numbers may not be that accurate. The craters left by the British WW-II Tall Boy bombs, 3'600kg of TNT equivalent, are notably larger than these predictions. See for instance the satellite images of the southern edge of the Håkøya Island near Tromsø, Norway, where the Tirpitz was sunk by 29 Tall Boys. Some of those bombs hit the island next to the ship. Note though, that these explosions happened underground due to the impact velocity of the bombs, which would create larger craters than predicted by the simple formula presented earlier.
Also, the explosion in Tianjin may have had the explosive material(s) spread out over a larger area, which could potentially create a much larger crater than predicted. Particularly over the wet, soft and sandy soil you'd expect in a harbour area.