The fact seems to be that it was a 4.2 magnitude seismic reading for an underground test- according to numbers I looked up 4.2 placed it in the few kiloton range. ER asked in that same thread if there was a lower limit to a nuclear bomb's explosive power- the answer is no- a sub-kiloton explosion is possible with a poorly designed bomb... or a VERY well designed one meant to be small.

The thing that makes a nuke hard to build- aside from getting the material- is taking a sub critical density of fissile material and quickly turning it into a super critical density (Note- the term 'critical mass' is a misnomer- it was reaching the critical density that triggered the Hirosima and Nagasaki bombs)

The bombs dropped on Hiroshima and Nagasaki assembled a critical density in very different ways. Little boy (dropped on hiroshima) fired, using high explosives, a sub critical U-235 'bullet' into a hollow sub-critical U-235 target to create a super-critical density -- this initiated the FAST runaway chain reaction that nearly immediately killed 70,000 men women and children, 15 kiloton equivalent. This is often referred to as a 'gun-type' bomb.

'Fat-Man' took a spherical 'pit' of subcritical plutonium, and surrounded it with a spherical shell of high explosive, when the shell was detonated (with a timing accuracy that pushed the limits of the technology then- since it had to be simultaneously detonated at many points at once to create a spherical implosion wave) the implosion wave hit the plutonium pit, the pit was compressed to reach critical density, and it detonated- 21 kiloton equivalent- immediately killing ~40,000 men women and children.

The key to a successful nuclear bomb, since 'success' is defined as 'ability to kill many people quickly', is to assemble a super-critical density as fast as possible- if you had the 'target' and 'bullet' used in little boy and tried to combine them using your hands they would melt and vaporize as you inserted the bullet into the target- you might get a little bang (and a whole lot of radiation) but you would not get a 'good' enough bang to kill 10's of thousands... similarly, if that gun in little boy had its aim off by a little it would also reduce its 'wonderful' 15 kiloton potential.

In the Fat-Man bomb, a slight glitch in detonation timing would lead to an asymmetrical implosion wave- this also would lead to 'disappointing' results...

If I had to guess, trying to make a medium sized (hiroshima, or nagasaki sized) bomb is easier than making an efficient sub kiloton fission bomb... but a small screw up could turn a nagasaki or hiroshima bomb into a sub-kiloton dud- extreme screw-ups lead to only a detonation of the conventional explosive and simply a scattering of the fissile material... (Keep in mind- I havent built one in a LONG time...)

so - in short there really is no limit on the lower side to the possible power of a poorly made nuke. My guess is that the North Koreans goofed and achieved a lower yield than they hoped. They woulld be shooting for the largest 'bang' they could... and making an efficient 'small' bang would probably be too difficult for them.

However, there ARE ways to make a large underground nuke look small:

http://www.nrdc.org/nuclear/acato.asp

A discussion of methods can be found half way down in the section entitled: How Probable Is Successful CTBT Evasion via Cavity "Decoupling"

but how likely is it- given past behavior- that KJI would want to intentionally make his bomb seem smaller?

I think they screwed it up, but it STILL is serious cause for concern....