Thyristors as an assembly often have snubber's across anode and cathode to protect them from dv/dt, as in voltage spokes due to inductive loads. The snubber is normally a capacitor less than 0.5 uF in value in series with a relatively low ohm 1 to 10 watt resistor. Capacitor is often rated 3 times line voltage, or more.
This causes a substantial leakage current to flow to the output stage at ALL times AC power is ON, perhaps 1 percent of total output current. Thus you may need a substantial load to bring the output voltage down close to zero, but it will never zero as long as power is ON. This will show up as voltage on the output, possibly enough to be a serious shock hazard or seem confusing to test equipment.
Proper testng means having a substantial load on the outputs, at least 50% of maximum load, preferably a resistive load to avoid back EMF. Leakage current in the snubbers and a light load can make "modes" seem irrational in their behavior.
This means that thyristor-based power controllers do leak enough current to be a shock hazard, especially if output has no load. In which case it appears to not shut off by using pulse-width control, as the snubbers can leak milliamps of current as long as AC power is ON. Best to work on these devices with AC power disconnected, and a minimum load to discharge snubber capacitors.
If the load is inductive then back EMF can be suppressed with series resistors of 5 ohms to 15 ohms WW at 300 watts, but wattage should be at least what the power controller can put out.