I'd want to do three or four tests to be reasonably sure:
0. Just ohm it out, of course. Make sure it's the polarity and type it's supposed to be, and not a FET or darlington or already dead.
1. Breakdown voltage. As you did correctly, just set it up as a zener diode. Obviously, having a higher voltage source comes in handy here, but you can make do with what you've got...
Don't need much current, just use a current limiting resistor or CCS. The datasheet gives what current it should be measured at.
2. Current capacity. Test Vceo(sat) at the highest current specified. This is usually near Ic(max), where hFE is dropping off quickly as well. You may want to set this up as a pulsed test, in which case, grab a 555 (yes, there are situations I will recommend 555s for... if not many
), put together some emitter followers to drive the base, and grab a low-value power resistor for collector load. If your bench PSU can't handle the pulsed current, put a huge cap across it as well (say, 4700uF+). Set the 555 for, say, 100us on, 10ms off, hook up the scope and take a look. Obviously, set base drive, collector supply and load resistance as needed.
This tests if the die is, in fact, as big as it should be. BJTs can only handle so much current density before hFE takes a huge dive (high level injection), and this sets the maximum collector current. If they scammed you with a 2N2222 die inside the power transistor package -- it'll fall over at less than an amp!
You can also figure out what hFE is at lower currents, which is important in determining if they gave you a high voltage switching transistor (low hFE) versus a linear one.
3. Power dissipation. This also verifies die size, and packaging quality. You can also test 2nd breakdown to ensure you've got the SOA required for an amplifier. This can be as simple as a common-emitter amplifier (base bias voltage, emitter resistor, collector load). Obviously, you need a power supply and heatsink big enough to handle all the power in question here (potentially over 100W).
Switching transistors have terrible SOA, so will be discovered easily, say at 50 or 100V and modest power levels.
4. Switching speed. You kind of have this already, if you opted to do #2 with a pulsed test! Just need one tweak: turn-off current. If base drive is supplied from a driver circuit through a series resistor, then you get turn-off current by using a complementary driver, and strapping an R+C across the series resistor. The C ensures that, even if your driver can't pull down to (or below) GND, turn-off current is supplied. This also gooses turn-on current. You can steer things around a bit with diodes, to get whatever turn-on and -off current you desire. (Just make sure not to dead-end the cap behind a diode -- if you do, put a modest resistance in parallel with the cap, so it leaks down during the off period.)
This helps determine if you've got a switching transistor (large t_stg, low hFE), or something ancient and slow (original 2N3055??), whereas you're probably expecting something moderately fast. You also don't want something too fast, lest it oscillate. Check against a known-good reference.
This should be pretty thorough, as far as testing everything a transistor needs to do. What it won't tell you is if they're pulling the same trick you're doing on yourself: authentic parts, removed from old equipment. That's much harder to determine. I'm not sure how exactly you'd figure that out. Possibly, increased leakage, or impaired hFE at low currents, maybe something at temperature extremes, hopefully not a lengthy accelerated-life test, or x-ray analysis!
Tim