19 Set restoration - Part 1: Power Supply

Having won my British WS19 MKIII on an internet auction site I was able to collect it person and check that it was as good as advertised before handing over my cash. Had this not been possible then I might have been reluctant to bid as high as I did.

However it came without its companion power supply, so having confirmed that the radio was indeed a suitable candidate for restoration, my next task was to find a suitable power supply to go with it.

The power supply that I eventually acquired is the PSU No.2 MKIII, Canadian version, which uses a uses a single dual-output dynamotor to produce +275V HT for receive and +500V for transmit. This differs from the British version which has two separate dynamotors, although both types are are interchangeable.

The Canadian version also has a receive-only 'vibrator' mode which uses a mechanical vibrator and transformer, and this usefully reduces the current drawn from the battery. It is also much quieter, although it is doubtful if this would have mattered much in a tank! In vibrator mode a relay starts up the dynamotor whenever the PTT is operated.

The power unit that I received had been quite badly hacked about and 'modified'. A length of 3-core cable emerged from a hole drilled in the front panel, with the remains of a UK-type mains socket on its end. Apparently this was a common post-war modification and some modified units were re-sold for use as inverters for caravans!

Top view showing relay and large capacitor

Most of the changes were easy to undo, apart for the relay from which several contacts had been removed. As can be seen from the photo, this relay is rather specialised and is possibly unique to the 19 Set, but surprisingly after only a few weeks searching I came across a New Old Stock (NOS) item from an internet seller. It wasn't cheap, but I decided that having a fully working supply was worth the outlay.

A rather special relay

After fitting the new relay, the next job was to replace the electrolytic capacitors. There are two 20uF 450V caps in a chassis mounted can adjacent to the rectifier valve.  I couldn't manage to remove this from the chassis, so I fitted replaceements under the chassis, hidden as best as I could. The large capacitor on top of the chassis was extracted from its cardboard tube. The end disc with the terminals was removed and the remains of the capacitor discarded. The leads of two new capacitors were passed through small holes made in the disc and soldered to the terminals. After adding some packing material to the tube (to prevent it it being crushed when the fixing clamp was tightened) the disc and capacitor assembly was glued back in place.

Re-stuffing the big 32uF capacitor

I replaced all but one of the paper capacitors, using mostly 100nF 600V polyester types. This was quite time consuming, as access to some of them was very difficult (and to one impossible!) Having struggled for a while I did manage to remove the front panel for better access, but all my attempts to further dismantle the unit failed. For this reason I was also unable to re-grease the dyno bearings, so instead I setted for running a few drops of oil into each one.

Side view - not much to see!

The cold cathode OZ4A rectifier valve was missing. I could have fitted slicon diodes in it place, but I was able to find a NOS replacement from an internet auction site quite cheaply.

The mechanical vibrator was also missing, so instead I installed one of the excellent solid state replacements available from http://www.royalsignals.org.uk/vibs/. I bought a ready built plug-in replacement for the original vibrator, but they do have part assembled options if you want to hide one in an original casing.

Finally the fuses were checked and replaced with the correct values.

The unit isn't completely silent in 'vibrator' mode as you can still hear a buzz from the transformer laminations, but the improvement in reliability and efficiency is well worth the cost of the solid state upgrade.

The job had taken quite a bit more time and money to restore than I had expected, but I think the result was worth it.

The next task was to obtain the so-called 'dog bone' cable that connects the power supply to the radio.

Wireless Set No.19

Keeping the blog updated during the summer months has always been difficult for me, and this year has been even more busy than usual.  I thought I'd take advantage of a bit of spare time over Christmas to bring the blog up to date

Just before Easter I started  work on my Wireless Set 19, which turned out to be my biggest project in several years.

WS19 in Sherman tank 1943 (National Army Museum Copyright)

The Wireless Set No 19 was designed originally for tanks and armoured fighting vehicles used by British and Canadian forces in WW2, but later also used extensively in jeeps, trucks and even as a man-portable ground station. There are many excellent sources of  information on the internet so I won't go into great detail here, but the WS19 provided two radios in one box, plus intercom facilities for the tank crew.  The 'A' set was for HF W/T or R/T over longer distances while the 'B' set was used for very short range VHF comms between tanks.

My first experience of the iconic '19 Set' was as a kid in the 1970s. Thousands of sets had been sold as army surplus in the 1960s and although I never owned one myself back then, many radio amateurs that I knew did. 19 sets were often heavily 'modified'  or even broken up for spares without too much thought, as they were cheap and pentiful. 40-odd years later 19 sets and accessories have now become very collectable and sell for big money on internet auction sites. I recently saw a complete (but not guaranteed working) WS19 station with an asking price of nearly £1000!

Sometime after the war, large numbers of 19 sets were completely rebuilt (without the 'B' set) and many were then put into storage. Markings on the front panel of my British WS19 MKIII suggest that it was rebuilt at the Newark workshops of the Royal Electrical & Mechanical Engineers (REME). Unfortunately the date has been rubbed off, but based on similar sets I've seen I expect that it would have been late 1950s.

My WS19 MKIII

In the next blog I'll describe what I had to do to get the set working again.

Keep listening!