The design is fatally flawed - there isn't enough margin between the minimum input voltage to avoid dropout and the absolute maximum input voltage of the regulators, so unless you grossly over-size the transformer to reduce the difference between its unloaded output voltage and its output voltage when loaded by a bridge rectifier passing your 500mA max. DC load current, you'll run into trouble: If the transformer voltage at full load is enough for 3V headroom for the regulators when the mains supply is at its lower limit , the unloaded output voltage will be high enough to destroy them.
The dropper resistors R1, R2, R3 make the situation even worse because the increase the differential between the max no load input voltage, and the fully loaded voltage at the trough of the ripple. However you may still need something there to reduce the regulator max dissipation at full load current. See below.
I recommend redesigning using LM317 and LM337 regulators for the main rails, as their 40V abs max input to output differential gives you enough margin for a 20V-0-20V transformer to be practical. For the capacitively coupled 48V phantom power supply, you'll need to stack it on top of the +unreg rail to get enough input voltage. However under high line, minimum load conditions there could then be up to 100V on its unreg rail so you cant use a LM317 regulator there. A
Supertex LR12 looks to be possible. If it doesn't provide enough current for the number of mikes you want to run, split up the mikes and use multiple regulators, one for each pair of mikes, or even one for each mike for better isolation. You'll probably want a transistor
capacitance multiplier circuit feeding the 48V phantom power regulator(s) to reduce the ripple and avoid hum breakthrough.
You may also want a capacitance multiplier feeding each main regulator to reduce ripple breakthrough - depending on how good a PSRR your preamps have. If you Zener clamp their bases at say 28V from the 0V rail, you can also use them as preregulators, to split the full load dissipation between them and the regulators. That would also limit the no-load input voltage to the regulators, allowing you to keep the LM7818 and LM7918 regulators you were originally proposing to use.
The average DC current you can draw without exceeding the transformer's RMS secondary current rating depends on the rectifier topology, and whether or not it directly feeds a reservoir capacitor. See Hammond (transformer division)
Design Guide for Rectifier Use.
You'll need at least a CT 20V-0-20V or 2x 20V, 35VA transformer for 500mA out, (with about 1.6VA margin) and more VA if you need a lot of phantom power current. The effective secondary voltage for calculating the phantom power contribution to the VA is 60V, so you'll need 1VA extra for each 10mA load at 48V.
To get enough headroom for the LM337 at full load current and minimum supply voltage you'll need at least 1500uF reservoir capacitance on each main rail. You can ease that requirement slightly by using Schottky 3A diodes in place of silicon 1A diodes in the bridge.
LTspice sim of 20V-0-20V 35VA transformer + bridge rectifier attached. N.B. the sim assumes transformer losses can be modelled as pure resistances transferred to the secondary. This isn't actually true except at zero load and the max rated load, but gives 'ballpark' figures that are near enough as long as you design with reasonable safety margins
Home
Help
Search
Login
Register
