5687 Linestage behaving odd...

[bibster]

Hi,

Couldn't get rid of this (spikey sounding) 50Hz in my system...
Managed to fetch a scope from my wife's school (For the weekend only, but hey..)

Measured a bit round in this thing, and found some 1v swing in the 6.3v heater supply for the 5687's.... Add some 2200uF to each one of the two 4700uF's in there, but no improvement... Well, these 5687's are sometimes even used with AC heating, and this swing looks like a rather clean sine to me, so this won't, most certainly, be the cause of my hum.

Then, while moving the probe around in the amp, and having it on high sensitivity, I've found the cause: The 5U4...

When i look at the output of the 5U4, it goes way back under 0v, and then in a straight line (spike) back up, little ripple, and some piece of sine, and over again.

Having the probe just near the 5U4, or the wire going to the choke, will induce this spike in the probe, thus showing on the scope.
Is this normal for a rectifier? To go way back under 0V....?
Having the 2nd channel of the scope on the output, these spikes are seen on the output as little straight lines, 50Hz... :-(

Anyway, looking at the datasheet, I couldn't find any info'bout that, but found some scheme's for PSU's and a hint saying to put no more than say 40 or 50 uF's in the first C after the 5U4.

I'v add one 47uF/350v right next to the tube, on its socket.
Spike is gone, bass is improved (as seen using RMAA), but lost a lot of harmonics
And having the probe near the tube won't induce any spikes anymore...
I've rendered this psu using psud2, but never seen this behaviour

My question: Is it a good thing(tm) to have added this C nest to the 5U4, and is it a normal behaviour for a tube rectifier to go way below ov, and to have this spike going straight up?
(I could try to upload some photos of the scope if that might help...)

regards, Paul

[bibster]

Noise graph

[bibster]

freq. graph

Rather strange, at the lower end, innit?