My new power supply

[(*steve*)]

I'm a sucker for an old piece of hardware...



A regulated power supply providing 3 rails:

1. 0-400V 100mA (B)
   
2. 0-100V 1mA (C)
   
3. 6.3 VAC 4A

It's nicely made, but pretty clearly not a commercial unit.

The meters measure either the B or the C output. However there is no deflection of the current meter on the C output (maybe because 1mA wouldn't show up, maybe because it doesn't switch it over). However there's plenty of room above the existing meters to add a smaller pair of meters, and maybe even some pots to add current limiting.



The case seems to be of some sort of bolt together construction, possibly also custom made.

Are you expecting to find a heatsink on the back?



Oooh, no... it has those vacuum tuby thing in it! Let's take a closer look.



That's a pair of 6L6GC's you can see there. At this point I don't know if they are in parallel or there is one for each output. These tubes are rated for an anode voltage of 500V and a forward current of 0.9A, so a single tube could handle the load. But why would you use one of these for 100V and 1mA?

A view from the top shows you the other valves



There's more evidence here that this is a custom or possibly home-made construction. The vintage of the transformer on the left doesn't seem to match some of the other stuff, but I may be wrong.

The two outer 7 pin valves are OA2's, and the middle one is a 6BH6.

The OA2 is a regulator tube with an operating voltage of 150V and current of 5mA to 30mA. Why are there two of these? And given the low current from the 100V output, would you really have a regulator tube drawing at least 5 times the maximum output current?

The 6BH6 is a pentode, and I guess this could be the output tube for the 100V supply, or it may be for something else.

Looking underneath, it's nice and clean:



You can really see this was finished off by hand -- look at the mounting holes for the two caps that are mounted on the upper side of the chassis.

Obviously there are lots of semiconductor diodes here, and apart from the construction, I think the thing that most dates it are the ELNA capacitors.

I have checked the basic performance of the outputs, both vary smoothly in the range expected, but the 6.3VAC is closer to 7.5VAC.

I have also loaded the 100V output to 1mA, and here things got surprising. The voltage sagged by about 20V when this load was applied. It may be that the output has a high impedance to protect the grids it may be connected to (if you make a mistake) or it may just be that the regulation is poor (or maybe non-existant).

I have yet to load-test the other outputs, but that's on my radar.

Also, I'm going to reverse engineer a schematic so I can see what's actually happening.

 

[davenn]

sweet, nice build

 

[(*steve*)]

After reading the specs for the 6LS6GCs, it's clear why there's two of them. They have a maximum anode dissipation of 30W.

 

[twister]

You sure that's a power supply? Could it be a audio amp?

 

[hevans1944]

Back in the day, its was quite a popular pastime to build a regulated power supply for experimenting with vacuum tube circuits. If you were messing around with power amplifiers, you would also need some negative grid bias, but not much current, for that. And the grid bias voltage should be variable, which was usually accommodated with just a potentiometer in parallel with the OA2 gas regulator tube. IIRC, typical values for the pot ranged from about 100 kΩ to 1 MegΩ, so you wouldn't expect good voltage regulation with a variable load. Well, anyway, that's how I did it. You want to bias the OA2 so it is drawing close to maximum current without any load. As the load resistance is decreased, the increase in current subtracts from the current through the gas discharge tube. With a pot connected across the OA2, and no load drawn from the negative bias supply, current through the OA2 should be set somewhere between 50% and 90% of its maximum rated current. This will allow a small margin to accommodate upward variations in power line (mains) potential without overloading the OA2.

Later, when things have evolved to the point of actually building something useful, the negative grid bias is usually accomplished with a cathode bias resistor, suitably by-passed with a capacitor.

The other OA2 was used as the reference potential for a series regulator employing a pair of parallel-connected 6L6GC (NOT 6LS6GC as mentioned in post #3) power pentodes as the series pass element. The 6BH6 was the "error amplifier" that controlled the grids of the power pentodes based on the difference between the output B+ supply voltage and the voltage set by the B+ adjustment potentiometer.

I used somewhat more robust pass tubes than the venerable 6L6s (which were popular audio speaker drivers), but was never able to drive my pass tubes into cutoff. Therefore my B+ power supply was only adjustable from about 200 volts upwards to about 600 volts or so with a maximum current of about 500 mA. This was more than enough range to power up an RCA 6146 beam power pentode that served as a Class C amplifier for my homebrew radio amateur (novice) transmitter. Voltage regulation was tolerable, IIRC less than 20 volts sag from zero current to full current, and from over-the-air reports there was never any problem with that. I used an RC shaped-keyer circuit to control the attack and decay of the CW transmitter output so as to avoid transmitting the dreaded clicks, and a separate crystal oscillator tube to avoid chirps (frequency pulling), both of which were a sure sign of a lid (inexperienced) operator.

Most hams in my day were already "appliance operators," preferring to use factory-built rigs instead of building their own. I actually built my radio receiver as a Heathkit since I knew I wasn't competent to design and build an amateur band receiver from scratch. I only managed to build the CW transmitter because it was (1) crystal-controlled in frequency and (2) operated only on the 80m Novice band. Plus, I was anxious to "get on the air" to try copying some Morse code amateur radio transmissions while waiting for my Novice "ticket" to arrive in the mail.

Steve, if that isn't a factory-built power supply, it was certainly assembled by a dedicated and highly skilled hobbyist. I used to marvel at the photos of "ham-built" equipment published in QST, only much later realizing how much effort went into making those projects look so damned professional. No real hams that I knew made anything that even faintly resembled the pictures published in QST. Your power supply would be an exception. Sure, their stuff worked, but it sure wasn't pretty!

Congratulations, you have a real treasure there even if you never use it for its intended purpose.

Hop - AC8NS

 

[(*steve*)]

Back in the day

I could quote almost everything Hop said. Hey, Hop, were you looking over my shoulder?

It's not anywhere near finished yet, and is hopelessly messy, but here is the schematic I've partially reverse engineered from the device.



This comes from the wiring under the chassis, there's a bit on the top that I haven't traced yet. Also, this is the first pass through, so there may be errors. Note that the main current path to the 400V output goes via the meter wiring on the top of the chassis, so don't go looking for that yet.

Note that the 100V supply is exactly as described by my prescient colleague.

All the interesting stuff happens in the cathode circuit of the two 6L6GC's (and yeah, I have no idea where that extra S came from). And that's very similar in some respects to a transistor used as a series pass element.

What is interesting is that the meter used to measure the output voltage has a 1mA movement. Throwing the switch to measure the output voltage must have a significant effect on the output voltage itself!

There's parts of this circuit that have me scratching my head. The two transformers, both supplying a high voltage supply make it a bit tricky. I think the power supply shown on the right of this diagram provides the power for the 400V 100mA output. The one shown on the left is for the -100V and the HT for the valves (and all the 6.3V windings).

 

[(*steve*)]

Measuring the output voltage of the 100V supply can cause it to drop by 20V (that's 13%!)

With the output turned all the way up, measuring the voltage using my multimeter only (151V)



And then switching over to the internal voltage measurement... (131V)