Itapirkanmaa2 builds a monster power supply:

I went back to my old hobby which is old radios.

This time around, I decided I need proper testing and other gear.

There's quite a bit of material that could be forthcoming on this field. I have accumulated lotsa stuff for "when I have time", but each of us only has some, eventually.

So let's begin with getting a proper power supply. Radios of course have one built in, in one way or another, but it may not always be functioning, so having one ready "off-line" is quite useful.

I could be buying a ton of cheap Chinese stuff for the power supply, BUT, and it's a really biggie BUT: they would be switching power supplies. Those will by necessity radiate extra radio frequency energy, and the suppression of those emissions may not be very effective, as the proper shielding costs money, which everyone in China would so much like to be saving. Moreover, working on radios, we do not want any of that excess radiation to mess up with our sensitive diagnostics & repair.

So, an analog one it will be. We can be sure that a power supply will take a lot of beating, and parts are liable to fly. I know that too well.. So it would have to be as foolproof as possible. 

Therefore, and for nostalgic & educational reasons, I chose a tube-based approach.

It's always easiest to emulate an existing design, so after some analytics, I chose a design that looks like it's coming from the German-speaking Europe from the titles and the typeface.

I also had a one given transformer I chose to use. The original had something else.

Specifically I have these secondary windings in the transformer;

500 volts 400mA

14 volts 3A

124 volts 30 mA

Since full wave rectifying would have elevated the voltage too high (about 500V *1.4), I had no choice but to go for half-wave rectifying. That will need considerably larger filtering capacitors, but I decided it's not a hurdle.

The thing is for any analog p.s. is that there will be losses most of the time and the losses will be the higher the lower the voltage given out. Naturally we want an adjustable voltage in the first place. This design goes down to 60 volts. The requirement for the auxiliary negative voltage stems from that lower limit, and this concerns the semiconductor analog power supplies as well.

There are two regulated voltages, the one supplying the control voltage to the set of tubes, and then the negative voltage for the low-side adjusting. They both are stabilized by Zener diodes (Zeners in series equal the sum on them all), and some standard semiconductor regulators on top of that (=high-side regulating).

The tubes I chose are 3 pcs of a US short wave/low VHF transmitter tube, known as the 7984. The predecessor was the 6146 transmitter tube, and the pre-predecessor was the great granddad 6L6 from the 1930s, which was a revolutionary and much copied design at the time, known as beam power tube (beam tetrode) https://en.wikipedia.org/wiki/6L6. The same basic 6L6 is still used today for nostalgic audio and instrument amplifiers. The 7984 is not in much demand, and can be had quite cheaply on eBay etc., until all the 6,6 million instrumentalists find out about it, of course..

The 7984 came very late in the tube era in the 1960s. It's made to fit into a 12-pin Compactron envelope, which was used a lot for the tubes of the 1960s TV sets, even into the colour TV era, and some radios. The Compactrons were hardly ever used in Europe, unlike in North America and the dominions, like Japan. One positive side is that there's no need for the awkward anode top connectors. The heater current is also "low", that is 850mA each, so my 3A winding will suffice. In fact I have spare current for the one other triode EC92. I don't want to lose any EC92s, so I'll use one half of an ECC85 which is the same tube doubled (as is the ECC81), but on 2nd thinking I don't want to lose an ECC85 either, so I'll take a PCC85 which was the version used in TV sets. That one has a 300mA heater, so in the end I'll have a whopping 150mA to spare in the current budget, and one bonus half of the PCC85, or one EC92, for tasks unknown, which could be adjustable current limiting. The heater/secondary voltages arrangement did require some creative thinking so stand by for what I came up with. There's also the matter of the anode voltage max limit, which is 550 volts peak and 300 volts permanent, so the ECC81/XCC85 should work fine there.

In the design I noticed something that many people seem to miss: there is a stated limit to the voltage difference between the heater and the cathode. In practice, old tubes many times begin leaking between the two as the physical difference is small. This will be a fault that can't be fixed and the tube needs to be replaced. The specified limit is around 200 volts typically. So here the heater will be connected riding on top of the power tubes' cathode circuit so that the heater vs. cathode voltage limit is not broken. Since all the secondary windings are well isolated from one another in the transformer, this is not a problem.

Much of the field can be observed from the schematics and my modifications. I'm using an electronic version of the choke, and saving a lot in bulky hardware. Why do we need a choke in the first place? To filter out the AC power residuals, assisting the large capacitors. Even more when there 's the half-wave rectifying at 50Hz. Several hundred mA is also not a light load when the voltages are this high. The choke is just a series-regulating FET and some simple components, I got one from a Dutch guy over on eBay, I had to beef up the components somewhat. 

The one regulating transistor (or its replacement) was pretty hard to come by, I needed more amps to be on the safer side, but all the actual high voltage transistors I had had a too low hFE for the original specification. I sat through a bunch of data and discovered I in fact had a suitable GE-12 which is a 300 volts transistor used in the Americas for AC powered portable record players and TVs etc back in the 1960s-70s, seems to me as a simple A-class amplifier. It's a silly TO-66 transistor case but I'm going to go without any heatsink for it for now.

Each of the three 7984s can handle 20W of continuous power dissipation and at least 30W momentarily, so I'm hoping for the best anti-destruction-wise even when the output voltage is set low. I'm going to use the self-resetting PTC fuses in the secondary instead of the traditional ones. In the original design there's a voltage halver switch to reduce the dissipation on low loads, utilizing the split winding, which I do not have here. 

You can see from the design modifications that I'm a very scared of any spurious HF emerging from the cadgetry. Therefore I'm trying to filter any mysterious spikes and oscillations out before the have a chance to set in during the operation. I'm also adding the RF filters to the AC and the DC outputs as per usual practice.

So while you're marvelling at the schematics, the legally required words of caution: DO NOT try this at home! There is enough Joules in the capacitors to send several people into heart fibrillations at once, at least if they join hands. See kaotic,com for more..!.

01-10-24 I got more 7984s from eBay. I had one already, and one Compactron socket for it.

I've ordered more sockets, but they are likely not here for a while yet. I'll probably find a way to convert my existing PCB version Compactron sockets to suit installing them "Manhattan style" on top of studs on the blank resin board. 

The 3mm thick resin board from Temu with the tubes, the electronic choke with the 1st stage smoothing caps that are likewise installed "Manhattan style", and the transformer,

Closeup of the transformer. 

More stuff taken together in Dec 2024:

The base is a resin board from Temu. They have 1-5 mm stock. The mains transformer will be accompanied by a small auxiliary one that converts the 124VAC into about 80VAC. I might extent the circuit by another 3 tubes some time. The tube sockets will be installed above the board "Manhattan style" with spacers.

One of the tubes has the flash mark on the side, not on the end, there's nothing wrong with it. There's the modified electronic choke circuit with the filtering caps that will probably need to be increased.

The series pass FET, the series pass transistor and the 7912 regulator are seen. A heatsink as well.

Some ferrite rings, I have 100s of smaller ferrite beads, somewhere to be found too.

A potentiometer with a power switch. A PCC85 tube with the socket. I´m going to use a microwave oven diode for the rectifier, which will also work as an anti-inrush current device.

TBC.