Vintage switched mode power supplies

[MrBungle]

That is quite amazing!

[Andrew2]

Originally Posted by kevinaston1

First one I ever met was the Thorn 3000

Likewise. The famous 'chopper' power supply! The idea was totally alien to me at the time and it caused some serious willies among myself and my fellow engineers.

[MrBungle]

Just had a look at the schematic for that and I'd be afraid of it! It's back to the old "stick beeswax in the grub screw holes to stop getting electrocuted" territory

Here for ref: https://www.radios-tv.co.uk/wp-conte...brc3k35k-4.jpg

[dazzlevision]

The export version of the Thorn 3000 PSU did use a double wound mains transformer.

[Biggles]

I seem to remember a Pye switch mode in-vehicle charger in the early eighties. I think the switch mode bit was to step up the 12 volts to maybe a higher voltage battery for a portable? I can't remember exactly but it wasn't easy to fix. Also Motorola made a 30A 12V power supply that was used on main site equipment which looked to me as if it was designed in the mid to late seventies. This had a conventional mains transformer and rectifier, but the regulation was by a couple of massive power transistors switched with a variable mark space. When under heavy load they used to audibly scream. When first switched on you could hear the oscillator start at low frequency then rise up as the regulator took over. It was always a good sign when you powered everything back up and could hear the whistling. Very reliable but it looked a bit lashed up inside. If the crowbar protection triggered they used to make a bit of a pop as the big fuse blew.
Alan.

[kalee20]

Originally Posted by wd40addict

The Bell & Howell 640 16mm cine projector from around 1957 uses a push pull pair of UL84s run straight off the rectified 110v mains input to supply most of the amplifier, see attachment.

With the main amplifier output powered by voltage-doubler rectified mains. Wonder why they went to all that trouble? Early stage heaters etc powered by HF AC would keep the hum down, ditto with exciter lamp. But deriving HT from rectified HF AC is a new one. Presumably the (?CuO or Se?) metal rectifier is inherently fast recovery? Not so elegant is that it's half-wave rectified from the HF transformer, so unequal loads on the UL84's.

Is this switch-mode? I'd argue not. I bet the UL84's are working in Class C (the anode circuit is tuned) rather than Class D or E, thus not dissipationless. But interesting all the same!

[kalee20]

Originally Posted by Hartley118

Bruel & Kjaer were early adopters of switch-mode supplies in their precision sound-level meters. Here's a section of the circuit of SLM type 2203 dating from 1960

Apart from a lesson in how NOT to draw a circuit diagram, this looks like a self-oscillating push-pull converter, with regulation achieved by a linear regulator controlling the push-pull stage's input voltage. So again, not switch-mode, but I can imagine it would be easier to shield and filter than a PWM switching thing.

[McMurdo]

I suppose 50's car radio vibrators were switch mode...literally.
The american space programme of mercury-gemini-apollo used switch mode power supplies in the spacecraft though I dont have many details.

[kalee20]

You could say they are switching, yes. And they also have lossless snubbing by virtue of a tuned transformer winding and associated capacitor which acts to reverse the voltage during the vibrator's dead time.

But they key thing that they don't do is regulate. A modern SMPS adjusts its output by controlling the switch 'on' and 'off' times, giving continuous adjustment of output, without any dissipation (save for the losses introduced by the non-idealness of the circuit elements).

[Radio Wrangler]

I used to have a Marconi H2900 receiver. A very early 1970s state-of-the-art HF rx developed in the late sixties. It had a switch-mode PSU regulated by pulse-width modulation.

Everything was in compartments in an immense cast chassis. No linear regs, if I remember, just huge amounts of filtering.

David

[ms660]

Syclops...Not that early but anyone remember them

Lawrence.

[Marconi_MPT4]

Interesting switch regulator circuit using a cold cathode trigger tube in Contamination Meter No.1. dated early to mid 1950's. The tube operates as a relaxation oscillator and switches to discharge C3. It was designed to regulate the battery supply over a fairly wide range.

Rich

[Ancient Geek]

The first one I came across was in the HP 5830A and 5840A gas chromatographs, the 5830 being introduced in 1974 amd HP's first processor-controlled GC. While most repairs were board swap, we had to repair these supplies down to component level in the field. Fortunately, by the time I was trained on them in 1979, the failure modes were fairly well known. The SMPS boards were huge and very complex by today's standards. There's an article about the 5840 here, if anyone is curious about what a gas chromatograph is:
http://www.hpl.hp.com/hpjournal/pdfs...Fs/1976-04.pdf

[Argus25]

Originally Posted by Radio Wrangler

I think refugee was referring to the field controller for the dynamo.

Yes, I misunderstood the remark, I was referring to the similarity of the Royer oscillator to radio vibrator supplies not the voltage regulator for the dynamo.

[cmjones01]

The EHT supply in Tektronix 500 series scopes must be a contender. It uses a valve oscillator driving a transformer at 20khz or so, and definitely regulates. From memory, it uses the screen grid for feedback. The design dares from at least the model 535 in the mid 1950s, and possibly earlier.

[Argus25]

Originally Posted by kalee20

True, although the field coil had a fair bit of inductance. So when the contact closed, the current ramped up; when the contact opened, the current ramped down. And the open and closed times were determined by a solenoid pull, with a carefully-adjusted return spring.

When I restore/repair the Lucas RB106 regulators I always add a snubber diode to the field circuit, without it the voltage can peak to about -70V when the contact opens and the arcing is objectionable for long term reliability. I have also added transistor assistance to the contact.

I have a large number of articles on these regulators, and electronic replacements, on my www.worldphaco.net website, how to repair & modify the RB106 regulators and how to make electronic replacements.

Also there, is currently the only published design of a "Dynamo & Alternator Emulator" which emulates the electromagnetic properties of a dynamo so that variable frequency regulators can be designed and tested say in a Spice simulation. These early types have an operating frequency that depends on the parameters of the machine. Later when LRC (Load Response Control) was required, these parameters became unimportant and the switching went to a fixed frequency PWM controller instead.

As a thought experiment, I wondered if it would have been possible for Lucas to have built an effective transistorized equivalent of the common RB106 regulator in the early 1960's. So I selected parts from that era and built a regulator which exactly emulates the features of Lucas's compensated RB106,You can see the switching waveform across the field coil at the end, the article is here, sorry about the crazy long link:

http://nebula.wsimg.com/16f83c1c70ef...&alloworigin=1

My formal replacement for Lucas's RB106 is in this article; I worked on the design for some years to make sure it was rugged enough for the application. There are many versions out there now but they contain fragile IC's and electrolytic caps and don't do well under the bonnet in the heat and electrically hostile environment. This article also explains the two forms of compensation that exist in Lucas's design and contain the equations for their regulator which were resolved using a specially built test machine:

http://nebula.wsimg.com/f167bd9d35e5...&alloworigin=1

[MrBungle]

Found a reference to transistor DC/DC converters from 1960: https://ia800706.us.archive.org/35/i...s1stEd1960.pdf

See chapter 26.

[Argus25]

What a wonderful manual, thank you very much for posting that. I have not seen that one before.

On page 139 is has one of my favorite amplifier circuits using an EF98 and a germanium power transistor and a 25:1 inter-stage transformer. The thing I have always liked about this design (despite the poor efficiency), it has an input impedance of 10 meg Ohm an output impedance of 3 Ohm, substantial power gain and it only has the two active devices. It took my fancy so much that it inspired me to build an entire radio using the EF98 and the OC16 as the output.

I have a circuit somewhere in a vintage Philips manual with a single ended switch mode supply, that I think used an OC77 and the date of it was around 1959 I think, so the switchmode concept goes way back.

There is also another type of switchmode supply used in aircraft, the Exciter. Initially these had a vibrating reed, but later they went to gas discharge tube. I could find so little information on these (partly because the designs are often proprietary and potted in rubber or resin) that I did up an article on how they work. I had a company in Shenzen make some GDT's for my experiments and I got a few GDT's from defunct exciters. The article is is here:

http://www.worldphaco.com/uploads/GA...R_EXCITERS.pdf

[G6Tanuki]

The mid-1960s all-transistor Racal RA217 communications receiver uses a SMPS (the switching component being a single NKT404 germanium transistor; another NKT404 is used - strapped as a power-diode - in the DC input to provide reverse-polarity protection).

The 'switcher' allows the entire radio to be run safely off a wide range of low-DC voltages while the supply to the radio's internals is regulated, smooth and stable . Just as important this makes the negative/positive-earth or floating-supply issue totally irrelevant because the SMPS provides full DC isolation between the supply and the radio's real workings.

The power-supply also caters for AC supplies: when running on AC the output from the mains transformer is rectified, half-heartedly-smoothed, and fed to the switcher's input.

[kalee20]

Originally Posted by Argus25

What a wonderful manual, thank you very much for posting that. I have not seen that one before.

Me neither. An excellent Mullard publication, with plenty of calculations.

The section on DC-DC converters is good, switching techniques to get voltage step-up. Although, there's no PWM so everything is unregulated.

Originally Posted by cmjones01

The EHT supply in Tektronix 500 series scopes must be a contender. It uses a valve oscillator driving a transformer at 20khz or so, and definitely regulates. From memory, it uses the screen grid for feedback. The design dares from at least the model 535 in the mid 1950s, and possibly earlier.

It does, and I was thinking about it when this thread kicked off! Although, it regulates by varying the screen-grid voltage, thus controlling the amplitude of oscillation. Essentially, it is a lossy technique, with the power valve operating as an amplifier rather than a Class D switch.