Here in thumbnail is a LTspice simulation of the circuit suggested in the previous post. The voltage output into the transformer is a 24-volt square wave at ~150 Hz. If a standard 60-Hz 120V to 12V (10:1 turns ratio) transformer is used, the 24-volt square wave represents about 2.2 times more volt-seconds at 60 Hz and would saturate the transformer core, so the LTspice circuit is running at ~150 Hz. This would increase the lamination losses (which scale as ω2) by a factor of ~6.3, so a standard 120V to 12V transformer is probably not appropriate.
The transformer in the simulation has a 1:10 pri-sec turn ratio with a center-tapped primary. The rectified dc output is about 360 volts into 10,000 ohms (36 milliamps).
More appropriate transformers would be a 120V-to-24V CT 60-Hz transformer (which would reduce the dc output to ~180V), or an audio output transformer with thinner laminations. In either case, a spice simulation and/or bench test is recommended.
[added] The waveform in the second thumbnail represents the first 250 ms of the Vac (unrectified) output.
(Request to moderator- Pls move to the EE forum. Thanks)
Bob S