the forward voltage drop
When a diode is conducting, it does not behave like a wire — it 'keeps' a little voltage for itself. That kept voltage is the forward voltage drop, written Vf. For an ordinary silicon diode passing a normal current it sits around 0.6 to 0.7 V. It is the toll the junction charges to let current through.
Crucially, Vf is not a constant of nature; it depends on current, and only weakly. The junction current rises exponentially with voltage, which flipped around means the voltage rises only logarithmically with current. So a diode might show 0.6 V at 1 mA, 0.7 V at 10 mA, and 0.75 V at 100 mA — ten times the current for barely a tenth of a volt more. That is why we get away with calling it 'about 0.7 V': over a wide current range it barely moves.
Two practical consequences. First, that 0.7 V times the current is real power turned into heat inside the diode, which matters in a 5 A rectifier. Second, because Vf shrinks as the diode warms (roughly -2 mV per degree C), you cannot reliably stack diodes to make a precise voltage, but you CAN use that predictable drift to sense temperature.
A 3 A bridge rectifier has two diodes conducting at once, each dropping ~0.9 V at that current. That is about 1.8 V lost, and 1.8 V times 3 A = 5.4 W of heat — enough to need a heat sink.
Vf times current is wasted heat; at high current it is not negligible.
Treating Vf as exactly 0.7 V is a modelling convenience, not a law. It drifts with current and falls with temperature; if your design breaks when Vf is 0.6 or 0.8 V, the design — not the diode — is the problem.