Quote Originally Posted by 2E0ILY View Post
Thanks to advice from here and on another forum I have made a decision. As I hope to build several amplifiers over a period of time I believe building a variable voltage output HV PSU first in a separate enclosure, would be the way to go. I have found an article in an old (1994) ARRL handbook showing variable voltage supply using a variable auto transformer on the primaries of a Peter Dahl Hyersil HV transformer that still seems to be listed by Harbach, and are hopefully still available to order.. I investigated having a transformer wound here in the UK, but few companies now exist here able to do this, and for a one off I have fond considerable disinterest. I may have to bite the bullet on shipping costs. The supply is to 3500 volts at 1 amp. There are two items causing me a little concern over where to acquire them. The circuit uses HV rectifier diodes from K2AW's Silicone Alley. They are HV14-1 part number, 14kV at 1 amp. I am not sure if K2AW are still in business, and if these are still available? I might have to source a suitable equivalent, any suggestions please? They also cite a 25uF 4kV oil filled smoothing cap, and here in the UK I can't imagine where I would find one. I am thinking of using an array of suitable electrolytics unless anyone knows of a source for a suitable oil filled device?

Given this supply I should have more than adequate power available for any reasonable project.
You're welcome. You don't have to use an oil-filled smoothing capacitor. Many use a stack of electrolytics (same mfr & pn, from the same batch, with individual bleeder/equalizing resistors). I suggest you stack some common 1N5408 (3A/1KV) or 6A10 (6A/1KV) rectifiers. The separate plate supply for my 160m 1500W amplifier will use 6A10s with a transformer I rewound, and a stack of ten 450V electrolytics. The calculations indicate 1.5A capability, with taps to provide up to 4KV (using a full wave doubler rectifier):



It came to me from a piece of medical equipment, with a split nylon bobbin (excellent primary-secondary isolation). Secondary DCR = ≈ 5.8Ω. LTSpice simulation with a 1A pulsed load indicates very good dynamic stability. Expected IP = 785mA. Thanks to XQ2FOD and others for info & design tools.