Puzzling Quench Results

[Buzzkill]

My younger brother requested a dagger.  I've never done that before so I opted to use some 3/4 inch coil spring for the stock.  I got some practice "uncorkscrewing" a blade  in the process, but overall I was fairly pleased with the results.  The bevels came out pretty well and I had no deep hammer marks to deal with so cleanup on the grinder was relatively minor.  After the rough grind it was time for 3 times of normalizing then quenching.  I use my propane forge with a 2.5 inch diameter piece of stainless pipe with one end hammered closed, so no direct flame impingement is possible.  Using my trusty magnet I went through my normal routine.  Heat to non-magnetic, back in the forge for 30 seconds, then into the oil (which I warmed already by heating and quenching another item).  I don't even have to take a full step between the forge and the vertical quench tank I'm using, so there's literally less than a second of time between forge and oil.  As usual I grab the file to check for hardness.  Surprisingly to me it bit in almost like mild steel.  So I heat up the blade again, but a little hotter this time and back into the oil.  The file still bites in.  One more time and still no joy.  So I grab the remaining piece of coil spring used for the blade, heat about 3 to 4 inches to non-magnetic and into the oil.  The file slides off like it was glass.   I tried quenching the blade one more time in oil and then in my frustration decided to do a water quench.  Same result.  The file still bites in.  A second water quench and the same thing again.  With this many quenches it's not surprising that the blade has developed a noticeable warp.  Since I'm not getting the blade to harden anyway I decide to straighten the blade cold.  Some of you probably already see where this is going.  With only moderate force on the blade it snaps the top 3 to 4 inches off cleanly.  I decided to make a shorter blade and ground a new tip on it, but leave it until the next day after that.  When I checked it the next day there were (not surprisingly since I didn't temper it) cracks near the tang area on both sides of the blade, which of course meant I had to start over.

So, I started testing the pieces of the junk blade a little more thoroughly with files.  It turns out there was a relatively thin layer of unhardened steel, but below it was fairly well hardened.  Of course my first thought was decarb layer, but this had already been through rough grinding after forging.  I thought that would remove any decarb layer. I run my forge rich to avoid scaling, which I also though would reduce the chances of that happening.  With the pipe in the forge and no direct flame impingement I figured any new decarb layer would be almost non-existent at that point.  Anyone got any idea of what was going on or what I should have been doing differently?

FWIW, the second blade came out pretty well, but did show the same symptoms after quenching.  However, armed with my previously gained experience, I only quenched once and got to the hard stuff quickly.  I'll try to get some pics up when I get done with it, whenever that is.