HT fuse located between center tap and chassis ground?

I just completed my Ceriatone OTS FM build (Dumble #102 clone), though I haven't really tested it at all yet, so there's still probably a lot left to do. But I have a question about strange behavior I'm seeing, and the HT fuse.

Here's the layout:



As you can see, the HT fuse is not like many other amps, in that it's not located between the rectifier diodes and the filter caps. Instead, it's located between the chassis ground and the high voltage winding center tap.

I kind of understand how this is supposed to work, in that current needs to be able to flow back to that winding in order to complete the rectified circuit. But I don't fully understand it.

The really weird thing is that when I switch on the power and standby on the amp, even if that fuse in not in place, the filter caps still start charging up, and I start to see DC voltage if I measure between, for example, chassis ground and the first or second filter cap.

Is this behavior incorrect? I sort of feel like it could make sense, maybe the transformer can maybe still produce voltage but not current? I'm not sure if I'm just excusing faulty behavior though. I mean, thinking about it another way, if the fuse can't stop the high voltage and turn off the amp, then what's even the point of it?
Can anyone advise? Thanks!

 

Hmm, I just found this:
https://el34world.com/Forum/index.php?topic=19440.msg201234#msg201234

Sounds like because I ended up using a different PT than Ceriatone intended, it's dangerous to leave the fuse on the center tap as is. The PT I ended up using has the bias tapped from the HT winding, not on its own winding, like in the diagram above. It sounds like if the fuse is blown or not inserted, the bias tap can end up serving as an unbalanced center tap, resulting in damage to the bias filter cap. I think tomorrow I will move the fuse to instead come between the diodes and the standby switch, it should be an easy change to make. And instead, I'll ground the HT center tap directly against the chassis.

I just hope I didn't damage the bias filter cap when I was measuring the voltage and experimenting. It seems okay, still reads 100uf on the multimeter, still does not read continuity between its two leads, doesn't show any burn marks, and is not leaking any fluids. So maybe it's not damaged...

 

I think that signal ground-chassis connection is best near the input jack, if I recall from http://www.valvewizard.co.uk/Grounding.pdf

 

Very interesting. But unfortunately, I don't think I have room for adding another fuse to have one on each leg. One, after the rectification will have to do for now, though maybe I could consider the inline fuses, which I did not know existed. But at least this amp has 3 diodes on each side, which I think should make complete failover of one side more unlikely.

Also wondering, if you go with the route of having two fuses, one on each leg, what should the rating be? Would it be half of the rating of the rectified fuse? So 250ma per leg?