My latest project of the week.
6/2/2002
This great looking little
set is the Ferguson. 366. 3 band ac/dc Superhet, built in 1935, was brought
to me about 10 weeks ago for a full 
restoration.
I was in a terrible mess. The carbon resistors and all the paper capacitors
were replaced as a matter of course, a new double pole on/off switch, It must
have been run for sometime on the incorrect mains voltage as 4 of the 5 valves
were low in emission. The smoothing capacitors were original and faulty, they
are built into an aluminum can. I find the best way to deal with this problem
is to neatly cut open the can remove all the gunge, then fit in modern capacitors,
after carefully refitting You would never know it had been repaired. The set
was designed to be used with a line cord. (A dropping resistor incorporated
in the mains lead, which supplies the power to the valve heater chain) This
method of reducing the AC voltage Is a bit of a fire hazard but very common
in the 40s and 50s as it saved the cost of a mains transformer. I'm sure it
would be regarded as unsafe to use these days. An alternative is to use a
large capacitor to reduce the voltage, and thanks to John Langley of Northhants
for his article in "Television" magazine which explained the method.
This system of reducing the mains voltage has the great advantage of not generating
any heat as it is reactance not resistance. The capacitor is shown in the
photo. It is the "Motor run" type and available from www.cpc.co.uk
This one is 4ufd. Its' value was established by trial and error bringing up
the mains voltage slowly with the variac. (Variable transformer). I had to
go back many years to get the formula for calculating reactance. It's an easy
matter
nowadays with a calculator to do the sum. So, the reactance in ohms of a 4
ufd capacitor at 50 hertz, I make it 790 ohms. ( Capacity in Farads)
Now comes the most important
part of a restoration, the assembly of all the refurbished parts, this is
where the attention to detail really shows. The refinished cabinet, the loudspeaker
fabric, control knobs backed with new felt finally the fitting of the dial
and the glass bezel. See below the finished set. After all this work was completed
and a realignment, It has turned into a real gem, with 
excellent
performance. The original cabinet finish and the loudspeaker cloth have been
retained. With very careful cleaning and touch in and then a final polish.
Since publishing this page I have received an e-mail from Paul Stenning, with a lot more information on the calculation of the value of the capacitor reactance. See below the extracts.
Hi Martin,
You only got half way with your capacitor calculations for your project of the week. Your formula for the impedance is correct, but it isn't as simple as using a capacitor having the same impedance (at 50Hz) as the resistor you are replacing. You have to consider the 90 degree phase shift too. If you measure the voltage across the capacitor and the voltage across the resistor (assuming the set is purely resistive, which for the greater part it is), then add the two together, you won't get the supply voltage, you'll get a higher voltage. This is due to the 90 degree phase shift. This is where some school math's comes into play! Right-angle triangles, "the square of the hypotenuse is equal to the sum of the square of the other two sides", and all that. Imagine a right angle triangle. The hypotenuse (long side that doesn't connect to the right angle) represents the supply voltage, and the other two sides represent the voltage across the capacitor and the voltage across the resistor. Now you can work out the voltage across the capacitor; you know the current through the capacitor (current drawn by the set) and from that you can work out the impedance required and hence (from your existing formula) the capacitor value needed. I get several e-mails about this, so I have written a piece describing it, which will go on my website shortly (once another techie friend has checked it for me), together with a spreadsheet to aid with the number-crunching. It's intended for filament dropper calculations (where the filament chain is purely resistive) but would still be somewhere near for a whole set because the filament circuit is still the largest part of the load
Paul.
For more on this topic go,