I spent some time chasing the 60Hz, I might hit a dead end.
1) I removed the input tube, very little change, the noise is from the EL34 to output transformer.
2) I shorted the input of the EL34 to ground, just remove a little bit of hash, no major change in the amplitude.
3) I removed the EL34 so only the output transformer. The waveform change, but the amplitude almost the same. So this eliminates the power tube, only the transformer.
4) I measure the center tap ( to +B) with scope on AC, less than 3mV of noise. But when I look at the plate connection( without the tube) I see 100mV. So is the screen grid tap.
5) I remove the red wire ( center tap) from the +B and solder to the screen grid tap, noise goes away.
6) If put a thick aluminum plate ( 1/8" thick) between the power transformer and the output transformer on the other side of the chassis, the amplitude of the noise gone down 20%.
7) I moved the wires of the output transformer, and other wires I can move, no change.
From this, it tells me that the transformer is talking to each other. I think it's the power transformer talking to the output transformer. The reasons are:
1) The center tap of the output transformer is quiet, noise did not come from the power supply. I see over 100mV on both the plate tap and screen grid tap, so it got to be picking up within the output transformer.
2) I suspect the core leakage flux of the power transformer coupled to the core of the output transformer. So I put a thick metal plate between the core of the power transformer and the output transformer. It lowered by 20%. This is very important observation, that the field is couple on the other side away from the circuit. Because it's low frequency, skin depth is very deep, I don't even think 1/8" of aluminum is thick enough, but it tell me the story.
I don't think packing the transformer this close is a good idea. The choke is almost right on top of the output transformer inside the chassis. At least the power transformer is orthogonal to the output transformer axis, so the coupling should be much less. The axis of the choke and the output transformer are in line with each other to give maximum coupling.
One has to be very careful in the transformer placement, once it's wrong, it is impossible to fix like in this case. The two output transformer already has a thick can cover them already, but it's just not nearly thick enough. Problem with circuit this simple, you reach the end of the road really fast and there's not much you can do about it.
I am open for suggestion and opinion. Please let me know.