Umm. Is this the one with the "reset" push-button on the front panel?
The power supply shown at
www.cbtricks.com/Amp/palomar/elite/palomar_elite500.pdf for the Palomar Elite 500 is going to be similar, even if it isn't exactly the same.
If you're getting about 19 Volts on the filter cap from the bridge, your problem lies in the regulator.
Only method that ever worked for me was to pull every transistor and check for failed junctions. The 5-Watt wirewound resistor on the emitter leg of each regulator transistor gets checked.
The chip is generically a "723" regulator. It's been made by a dozen different vendors since the original Fairchild version called the "uA723" was introduced around 1970 or so. The manufacturer will put a different prefix, like "LM723" for one made by National Seminconductor, or "MC1723", if made by Motorola.
If the chip is bad, it's a near guarantee that other parts that it controls have failed, and overloaded the chip. I insist on using a 14-pin socket in the circuit board any time this part gets replaced. The foil traces will usually tolerate the stress of replacing the chip once, before they lift up from the board. Using a socket means you only have to unsolder 14 legs on that part ONCE. If the chip pops again, replacing the next one is not so big a deal. Less labor, less stress on the circuit-board foils.
I'm guessing that the regulator uses a quantity of TIP35 flat plastic power transistors. If only one or two goes bad, mixing and matching brands may not work well. It's not really necessary to use a "matched set" of those, but making them all the same brand (and if possible same batch #) is usually good enough. Replacing all 3 or 4, (or however many this one has) at once is more reliable even if only one or two has failed.
Troubleshooting a power supply from voltage measurements can be tricky. Checking the parts first before applying power tends to work better. The 723 chip is the one part you can't easily check on its own. But so long as all the transistors and resistors check okay, then it's time to blame the chip if it still won't deliver power to the linear side.
73