WARNING: This project is connected to mains line voltage during its entire operation. This project should NOT be attempted by anyone who is not confident, nor competent around mains voltage. This circuit is potentially dangerous, if not fatal if accidental contact is made between you and it. Please adhere to basic electrical safety at ALL times and do not attempt to build this if you don't know what you're doing. All incoming, and outgoing electrical connections should be insulated, and the circuit itself wired using double insulated cabling. For safety reasons, this circuit should be mounted to a metal chassis with a plastic cover over the top of it to prevent accidental contact.

Toroidal transformers larger than 300VA present a large inrush current (some times in the order of 115 amps!) for a few milliseconds at power on. This can lead to nuisance blowing of fuses, or tripping of circuit breakers. A "soft-start" circuit helps limit that inrush current for a few cycles of the mains voltage, allowing enough time for the transformer/PSU to settle down, then allowing the full line voltage and current to the transformer. E-L core transformers also present an inrush current, however not as much as the toroid counter-parts (of similar VA rating). Even some high current switch-mode power supplies can present large inrush currents, as well.

This circuit gives us such a soft-start delay, and comes from an article published on Rod Elliott's ESP articles pages.

Link to Rod's article where the circuit comes from:

https://sound-au.com/articles/soft-start.htm

Figure 6A is what the PCB is based on, minus the 1A fuse. Having said that, this circuit should have a line fuse in series with the incoming active connection to protect both the circuit and the transformer. Use the recommended rated fuse by the manufacturer of the transformer, always! Please read the article thoroughly if you are in a 110-120VAC 60Hz country, as some component values are different for that voltage.

The 5W 150 ohm resistors should be mounted on the copper side of the PCB, with their leads insulated, so they come in to contact with the metal chassis with a drop of thermal transfer compound. The photo shows them mounted on the component side only for clarity.

For some reason, I made C2 470uF/63V, this should really be a 100uF/63V as per the circuit in the article. This will give a shorter delay of around 100ms, as compared to 280ms with the 470uF. This will also provide a faster reset time using the lower value capacitor. In my circuit, with a 470uF, the reset time is around 500ms, which is borderline too long.

And here's a video on building the project: