Core32 wrote:Yes, I already have a matched pair of 5U4GB tubes and I really want to do an all-tube design. (A pair of mono-blocks.)
Probably because of the challenge as much as anything.
My first two quarters of EE classes I had tube theory and design along with DC analysis.
If you're gonna go that route, you have to remember that the biggest differences between PS design with hollow state diodes like the 5U4 is this:
much larger forward voltage, much smaller Isurge ratings. You can get the spec sheet for the 5U4 from
Frank's with the plate characteristics.
Even though specced for an Isurge= 1.0A/plate, that isn't as large as it seems at first, and can easily be busted, especially with a C-in filter. You can also be looking at a Vf= ~100V -- whereas a Si diode drops a volt or less.
That was one problem I had with one of my designs: a PTX with an ideal current rating, but it overvolted badly with Si diodes (458Vdc, whereas I needed 350Vdc design nominal). I had to lose 100V, and considered a MOSFET dropper in the negative rail. Since this PTX also had a 5.0V/3.0A secondary, that pretty much said "5U4GB". Running the calcs with Isurge= 0.8A, gave a Vf= 100V. It was pretty obvious this PTX was designed to produce 350Vdc with a 5U4GB. That's what I needed, and that's what I got.
The other problem is excessive reservoir capacitance. A 47uF capacitor busted the Isurge spec, and needed to be dropped to 34uF (two 68uF/300V in series with voltage balancing resistors). That'll cost a few more volts due to averaging. All problems you can avoid by going solid state, and including a really big capacitor.