coppatone GM70 /6

OK, let’s get the power supplies on the map.
Looking back at jc’s DA100 amp, there are these traits I like to preserve. I quote from that post: ‘some well-known jc tricks, like tube diodes that provide slow turn-on of the voltages and dampen the spikes of solid state diodes; chokes in the ground leg of the pi filter. a new trick is chokes before the rectifier: AC chokes, no air gap needed.’

A 840V full-wave bridge supply for the GM70 should look quite familiar:

Only change I really made was a single TV damper (6cj3 can take a lot of current) and putting the final choke in both legs but on a common core. Concrete values for chokes and caps have to follow in a later post. It takes some real modelling to integrate the PS with the OPT and the GM70.

If a power transformer with much higher voltages is available, then a full-wave supply can be build:

No need in that case to include solid-state rectifiers; two dampers will do a great job. This supply will need a secondary voltage in excess of 1320V (ct). The bridge variant above half of that (more than 660V).

Speaking of the bridge, ‘BR’ in the first drawing does not stand for ‘bridge rectifier,’ it means Buddha rectifier:

I picked this to introduce a touch of (low noise) sophistication. It is swiped straight from the work of the late John Camille. Combine these VSAC workshop notes with the Buddhafied Afterglow article in this issue of Valve magazine and you got the basic idea. Two diodes in series are really necessary at these voltages, especially in the circuit below. The caps are ceramic discs. The rating of the resistors depends. For a 120mA mono supply, half a Watt is derated (a factor of four) enough. When building one supply for both channels, it goes up quadratic, so 2W resistors must be used.

The input, driver, follower and bias stages need a +/-500V supply. A while ago I thought we could do something tricky and stick the chokes and damper in the common leg:

This does not work. If in the 3-line system like this the impedance is in only one line, the bulk of the current will find a way to run through the other two, unimpeded (unless one can force all current through a common leg).

So everything is there twice and independently after the Buddha rectifier:

The final chokes of the + and - supplies are separate items. All the AC chokes, before the rectifiers, in all the supplies shown here are separate items too, lots of AC Volts going on there. The +/-500V supply cannot be created from the same secondary winding as the 840V supply (needs galvanic separation). Also here concrete values for components will have to wait for another post.

For all the filaments that we have to light, I am looking at the Rod Coleman filament supply kits/boards. These are really affordable, have a great reputation sonically, can be set up for any tube, including GM70 and they come with a schematic. Have to talk to my friend Daan about these, he has already got them.

That’s it for today, more details next time.