Are you measuring the voltages with or without load on the supply? Nothing plugged onto J3 = no load. Something plugged onto J3 = loaded (unless cables are disconnected at other end).
You could also have a meter that's measuring Vpeak instead of Vdc. Majority of digital meters out there will measure Vpeak if there is no real load on the supply.
To curb things a bit - *temporarily* tie a load resistor between E1 and TP3 using aligator clips. Do you have something like a 4.7K 5W resistor handy? See if that'll bring things more in line. If so, then it's a non-issue and we're just chasing our tails with the high E1 voltage.
Not sure I would mess with this, though. Just replace CR5 and give it a shot.
A better replacement for CR5 would be a 1N5372B.
Original CR5 was a 1N4759A which is a 1 watt part. During normal operation - that part continuously dissipated 0.715 Watts ... 71% of rated load => HOT! The 1N5372B is a 5W part. Larger physical size and will stay cooler.
When I do new boards - I use 1.8K resistors paired with 1N5372B's to keep heat down. I try to keep part loading at 10% or less to keep things cooler. This would bring the 1N5372B's dissipation down to 0.51W on a 5W part.