This PS should work as intended, right?
I am building a balanced amplifier that takes a single voltage input (onboard railsplitter and buffered vgnd), so was planning to use a slight variation on Carlos Filipe's Snubbed LM338 design, but with an onboard buffered railsplitter to provide the +/-18v needed by the DRV134 BLDs I have (digi01 boards). I am going to be using the 250VA 30v+30v Avel toroidal transformer from Parts Express, with the secondaries paralleled, which I will then rectify, and have 2 regulators sitting on that, the snubbed LM338, and a basic LM317 as shown in the datasheet to bring the rectified dc voltage (should be ~42v-diode losses, so about 40.5v or so, right?) down to exactly 36v, and then split that with a buffered TLE2426 for the BLDs, while the main amps uses the output of the LM338.
For the most part, I am pretty confident that this will work, but the one thing I am unsure of is whether the virtual ground will be stable, or would it require a seperate transformer? It seems like it should work to me, but I havnt done anything like this ever before, and cant find reference to something like this either. The voltage from the LM317 will be a fixed 36vdc, while the output of the LM338 will be usually 36vdc, but I will probably have it adjustable via trim pot instead of using fixed values.
I realize that the above was a bit garbled, but hopefully the schematic and board layout is enough for someone to tell if it will work to give out 36v and +/-18v.