I have a tip122 and discovered my multimeter uses Vce=2,8V DC and Ib=10uA for its hFE test. When I insert the transistor on the terminals for hFE test, the gain on the multimeter alternates between 1 and 2 when by datasheet information it should be 1000. Anyone knows why?
Added to get this image in here.. the nominal resistor values vary significantly between suppliers- the below is from a Motorola model with 8K & 120 ohms.
As Andy's answer suggests the problem is in those danged resistors. With 10uA there's only 70mV across the B-E junction of the input transistor and 0.7mV across the other B-E junction. They're both 'off', the "base" current flows out through the "emitter" lead, so hFE = 1.0.
Just for interest, here's a plot of hFE vs. base current. There is a similar shape to the collector current curve in this datasheet, but extended to low currents.
And, finally, to compare with the datasheet curve, here is a simulation of hFE vs. collector current (comparable to the datasheet plot). As you can see, the apparent straight line (on a log-log graph) decline from the datasheet can't be extrapolated very far.
That said, those "danged" resistors are there for a very important reason. Without them, even small leakage (especially important when the transistor is hot) would be amplified by the product of the two hFEs and could cause problematic leakage in the compound device.