Incorrect.......the depth of hardening is ultimately limited by the inability of carbon to diffuse deeply into solid steel.....check with any metal laboratory or metallurgist.....
Not a metallurgist, But, only partially correct I believe. The penetration of the carbon does indeed slow the deeper it penetrates & for practical purposes there is indeed a limit of depth to be reached. However, I do believe that on a thin structural part, as for instance the side walls of a Light-weight Flues frame it would be entirely possible to achieve "Too Deep" a case leaving a part subject to cracking from its brittleness.
I have one question which I can't recall ever seeing addressed. When the part was originally Carburized it was brought up to heat in a carbon rich atmosphere & held at heat until the desired depth of penetration was achieved. Annealing the part removes only the hardening effect of the subsequent quench, it doesn't remove the carbon. Now upon subjecting the part to Re-Carburization does it immediately begin to obtain deeper penetration.
From my limited knowledge of the subject as I understand it the critical temp of a piece of carbon steel varies according to the % of the carbon, the higher the temp the lower the C'temp. It would thus seem to me a high enough temp would need to be maintained prior to the quench to harden the entire depth of the case. It would thus seem quite plausible that at some of the very low heats from which some quench to help in the prevention of warpage they might thus be achieving a very Superficial hard case with the major depth to which a declining % of carbon has attained might well remain unhardened. While this might seem desirable to minimize wrpage, it would also seem to be less than desirable for the purpose of hardening.
While true I am not a metallurgist, I did spend 35 years in machine shops where heat-treating of various types (including case-hardening, just not for color) was carried out. My experience was that any time a piece of steel was heated to above its critical tamp & quenched there was a modicum of
RISK involved & that Every Time it was subsequently Re-Done, that risk grew slightly.