The only thing with using 'junkyard steels' is you have to be willing to sacrifice a knife or two.
If you can harden a test piece like was described earlier, go ahead and make a knife. After you quench it, draw the temper back at around 350 degrees, put an edge on it and abuse the knife. Hack away on a 2x4 and check the edge - if it chipped the blade is still too hard and brittle, if it bent it's too soft. The first problem is easily fixed by tempering at a higher temp (say 400) the latter problem means the knife needs to be normalized, rehardened and tempered at a lower temp.
If there are no problems with the blade after chopping on a 2x4 do the brass rod test Rob described somewhere. On a properly tempered knife the edge will not chip, crack or be permanently deformed. There are some exceptions to that test however (you can get away with a harder edge on a small knife that won't see hard use - like a small caper or a neck knife).
The majority of common steels you'll come across that'll make a decent knife can be quenched in oil - most any oil actually. I've been using 1095 which has a pretty 'hard nose' for a simple steel (you need to drop it's temp quick for it to harden) and if you look at its spec sheet, it's listed as a water-hardening steel. I'm quenching in corn oil and I *know* it's working well. Used motor oil didn't seem to work as well - it still hardened, but the quenched steel didn't seem quite as hard. Water worked extremely well, but if you (being ME) overheated the steel even slightly there was a good chance the blade would crack.
Though I mostly use known steels, I have to agree with Frank. Using reclaimed steels (resurrected
) appeals to my, uh, frugal nature and the blades do have something special about them.
Here's a plug for Frank's knives:
I have one of his sawmill knives. The first thing I did with it was put a scary sharp edge on it and hack most of the way through a doug fir 2x4. It still shaved hair after and there was no damage to the edge. The brass rod didn't phase it either. Would it have passed the ABS JS test - well, no, but it certainly wouldn't be due to edge retention. :D