This got me thinking about how easy - or otherwise - it might be. I doubt anybody has a practical use for such a device these days, so I present it as historical curiosity.Terrykc wrote:I first came across this idea very many years ago when someone published a low voltage transformer design which, by altering the order of several secondaries and connecting them in or out of phase, could derive a wide range of output voltages adjustable in one volt steps!
Obviously we need to start with a one volt secondary to give ±1V steps. The next one will be three volts which will provide outputs of 2, 3 or 4 volts dependant on whether the 1V winding is connected and in which phase. For five volts, if we add 5 and 4 we get nine so, conversely, we can subtract the first two windings from a 9V winding to get 5V and work the precious sequence backwards to get 1V steps up to 8 then, after 9, start adding them again until we reach 9 + 4 = 13V.
Continuing this process gives a massive jump for the next secondary to 27 volts! When added to the other nine, three and one volt secondaries in phase we reach a grand total of forty volts and can now generate any voltage up to 40 in one volt steps. The table below shows the sequence:
|1||1||9 + 3 - 1||27 - 9 + 3||27 + 3 + 1|
|2||3 - 1||9 + 3||27 - 9 + 3 + 1||27 + 9 - 3 - 1|
|3||3||9 + 3 + 1||27 - 3 - 1||27 + 9 - 3|
|4||3 + 1||27 - 9 - 3 - 1||27 - 3||27 + 9 - 3 + 1|
|5||9 - 3 - 1||27 - 9 - 3||27 - 3 + 1||27 + 9 - 8|
|6||9 - 3||27 - 9 - 3 + 1||27 - 1||27 + 9|
|7||9 - 3 + 1||27 - 9 - 1||27||27 + 9 + 1|
|8||9 - 1||27 - 9||27 + 1||27 + 9 + 3 - 1|
|9||9||27 - 9 + 1||27 + 3 - 1||27 + 9 + 3|
|10||9 + 1||27 - 9 + 3 - 1||27 + 3||27 + 9 + 3 + 1|
From memory, the original article suggested taking a battery charger transformer of suitable rating, dismantling it and removing the secondary winding(s), counting the turns at the same time in order to derive a figure for the turns per volt - typically four.
Then wind on the four new secondaries and reassemble.
What eludes me at this point is how the output voltages were selected. This was essentially a hardware project but I don't think a 40-way rotary switch would have been an option!
The only thing that I can come up with is to use eight DPDT toggle switches and a look-up table like the one above. The configuration is like this for each secondary:
One switch determines whether that particular secondary is in use or not and the other selects in/anti phase connection.
If anybody can come up with a simpler solution I would be interested to see it!