Still going to do a choke input supply, so the cathode(s) of the rectifiers gets attached to the choke( as drawn in the original schematic).
here is a discussion on the subject: http://www.tpub.com/neets/book7/27b.htm

it shows two SS diodes, the banded end is the cathode, in a vacuum diode, this is the hot part, either directly, as in a 5U4, or indirectly as in a 6CA4.

keep asking...the link to ducnan amps has a link for tube data sheets, and a downloadable program to simulate pwoer supply behaviour.

regards,
Douglas

• data sheet and other good stuff.

Any ides what the cost of the choke is? I'd like to have a ballpark figure, just so I can get ready for sticker-shock.

I downloaded PSU Designer a few weeks back and have been playing with it. Pretty cool.....Colin

For construction ease, a step drill is easy to get that can make a proper sized hole for a 9-pin. A hole saw or hole punch for the octals are the two simplest means.

If I had to pick one, I'd say 6CA4. If I weren't slightly over the voltage ratings for this tough little valve, I'd use it for rev. 5.

The 8039 choke is ~$75. If you are willing to limit yourself to 10 mA per chanel maximum draw, and not do the shunt reg, you can get by with the little Hammond. The 8039 will leave you room to expand, and will work quite well all the way down to about 10 mA of load, which is ~half the design goal of 10 mA per channel.
regards,
Douglas

If you have a copy of OrCad, you can click the image above and download the SCH datafile. If you make changes to the circuit, new mods or whatever, send them to me and I'll upload them and make them available to everyone. This makes a standard format that's easy to use.

If you do not have OrCad, you might be interested in the demo version. Click here.

I drew this schematic quickly and didn't include the power supply. It isn't much trouble at all, I just have been really pressed for time this week. I'll try to add that later, and that wil let us decide if we're going for the dual-diode full-wave rectifier.

Another thing that could be improved is that the proper values for the active components could be used. Right now, the schematic is right but the electrical parameters aren't. As an example, I grabbed a small signal triode but not a 5687 so the tube characteristics would need to be entered for PSpice to work. The diodes are 1N914's or something like that, which aren't the same as LED's. So if we want PSpice to work, we'll need to find the right models for each component or create our own.