I have a DIY box ( with Lundahl LL1620's inside), with multiple bias voltages. The voltages are taken off a zener diode string so should be relatively accurate.
The effect of different bias is easily auditable but not huge. There is not that much difference between 580 and 620v.
If you just want to try a different voltage, the easiest way is probably by using a bias-adj box, this box has a pack of battery inside to provide the differential voltage. Say, some 9v batteries in series, 3x will provide around 30v to be added/subtracted in the bias line.
The proposed bias-adj box:
--- box has a short input pig-tail which plugs into a stax-competible amp.
-- inside the box, signal lines go straight to the output socket.
-- The bias line from the input pigtail goes to the battery pack; the other end of the battery pack goes to the bias pin of the output socket. Change the # of 9v batteries to get higher/lower final bias voltage at the output socket. The batteries will last basically forever since there is no current drain.
switches can be added to pick which differential voltage (say, 2x or 4x 9v) or change the direction (add or subtract from the amp's bias voltage).