We need some dialogue regarding this potentially dangerous issue.

I just want to state that this is true for any CCS load; not just Damir's design which is designed safely well within the proper guidelines.

I would not want to think I had to always stay in the room whenever a CCS amp is on in order to keep watch on it.

Reading this, someone could think that there`re dangerous terrorists here... Everything can be destroyed and can be dangerous. High voltage can kill you. You can burn your finger on soldering iron. Loud music can damage your hearing.
Drain to Source and Drain to Gate max. specifications of DN2540N5 is 400V. We used B+=440V, but voltage drop through the CCS (6C45Pi example) is 440-150=290V. Counting the negative peak of the driver, 290+75 = 365V. Yes, it is safer to lower the B+ to 390V with 6C45Pi.

Another "horror" possibility is that somehow driver tube shorts, and that we have the ground on the CCS output. Or for some reason, CCS shorts and we have 440V on its output - then driver tube would "pull" large current, especially if biased with LEDs or Zener "under" cathode.
We can use, non-flammable (say metal-film) resistor of, say, 3k3/0,5W between the CCS and the tube. It would limit the current, and in extreme cases burn out (without flames), (hopefully:-) protecting the tube. I`ll try it these days - with 9mA current it dissipates 0,27W, and max. dissipation of 0,5W reaches with 12,3mA through it.
300B output tube can`t be damaged in those cases - it is connected to the driver/CCS through the big capacitor, in my case Mundorf Supreme, 4µ7/800V.
Adding "back-to-back" Zener G-S protection is another idea.

For now, I used DN2540N5 for about a month, few hours a day, with numerous in-out switching and other torturing with no problems and good sound. Yes, I burned one mosfet accidentaly shorting it with unisolated wire. No big deal

Huh...if I`d think that way, I`d never leave the house or sit in the car, or tinkering with tubes...hey I measured today 390-0-390 Vrms on the PT secondary - that`s 780V AC between those two red wires on my floor...and 440-460V DC besides...but I stayed cool...(I mean, no fire)

On more serious level, I tried 3k3 resistor today, 5-minutes job. It`s about 33V voltage drop "across" it, and less voltage "across" the CCS (440-183 = 257V). For feeling even more better, drop the 440V to 390V through RC filter, 5kOhms/2W - 47µF/500V. Now you have only 207V "through" the CCS, fine.
I listened to the added 3k3 "version", and without this resistor, and can`t hear the sound difference - good.

In the next few days I`ll post another "report", 6SN7 "family"...

I want to add that I only bring this issue up with you because you are willing to discuss it; this in no way is a criticism or knock on your excellent design. Using these CCS in Guinevere and now in a Tube amp results in a total of four CCS loads in use whenever I turn on my system. Thats a lot of solid state loads operating there.

If you read John Broskies analysis of the Curry Cascode triode amp with CCS and current mirror loads he says regarding the use of SS components stated at the end of the article;
"You must choose either safety or sound."

• http://www.audioroundtable.com/GroupBuild/Projects/Guinevere_Schematic.gif

However; I can't find any examples of failures or malfunctions by searching the net; so I guess they are pretty safe in use. I am sure if there were an event it would be written up somewhere on the net; so if there is I can't find it.

Better safe than sorry. Thats the only reason I asked.

No really...there is not much to go wrong. As Damir pointed out, MOSFET failures do happen, and I have had a few, in several circuits, and I have *NEVER* had to do anything but replace the FET's.
cheers,
Douglas