What is the required VA rating of the transformer in this LTSpice power supply?

Question

I have this circuit that I'm going to build, including the transformer. It is a simple power supply, with RC filters and one choke, I know all the voltages and currents.

Because I will design, wind and build the transformer myself (I already know all of that as well) I need to know the VAs that it must withstand.

I know that the total current load pulled from the transformer will be more than just the sum of all the DC sources because the capacitors pull reactive power that is not "used" by the circuit itself but just shuffled back and forth to keep the voltages at the nodes clean and constant.

I also know from past experience, that to find the VAs, it is not as simple as just multiplying the rms AC voltage (green, 358VAC) and the rms current flowing through the transformer (red, 1.21A) because the shape of the waveform as well as phase has a lot to do with it. The AC voltage is a sinewave, but the AC current flowing through the winding of the transformer (L2) has spikes. For example, if I just multiply both values I get ~433VA which is a lot more than I would expect.

So, the solution is to ask LTSpice to do the work for me and ask it to plot a graph of the power consumed (pink). The problem is that it shows the average value, not the rms, and there is the dilemma. Are those 144W average the VAs that the transformer will be working with and thus, the figure I have to design the transformer around? If not, how can I calculate or approximate the RMS value from that AVG value?

EDIT: This circuit is the power supply for a 50W vacuum tube amplifier I'm designing. The power consumption of a commercial amplifier that puts out 120W is claimed to be 300W, so the reading LTSpice is giving me does make sense, it is a little higher still considering that the transformer will have more windings that will draw an extra of about 35W in total, but that's because efficiency overall decreases with power, so the 144W is in the ballpark, I just want to confirm it.

Answer 1

You don't want to know the VA of the transformer, at least not as a single figure.

The core details and your working flux level will give you the turns per volt and the core heating, which are more or less independent of current.

You want to know the load rms current, which gives you the copper heating. Modify your Spice simulation slightly to compute the average current squared when driving your rectifier.

If you multiply working voltage by rms current, then you get a VA figure, but it's less useful than the two figures that contribute to it.

Answer 2

I think the main problem is that your simulated transformer has unrealistically low winding resistances. This results in high rms current because the winding resistance is not limiting peak rectifier current sufficiently.

I simulated a simplified version of your circuit in LTspice. Using primary and secondary resistances of 12Ω (a guess based on a 150VA transformer I measured) the rms current dropped from 1.2A to 590mA. At 358V this equates to 211VA.

The higher the VA rating of the transformer the lower its winding resistances should be. If you used a 400VA transformer the rms current would be higher than with a 200VA transformer. Therefore your simulation should include expected resistances for the size of transformer you intend to use, rather than the extremely low resistances of an 'ideal' part. Where possible, try to use accurate models of actual components for all devices (including diodes and capacitors), as these also have resistance which can affect the simulation.