# Why need voltage doubler when transformer can be made for full voltage

I have seen old vintage tube power supply (for example Heath HP-23A), that transformer secondary is roughly 300v RMS and a voltage doubler is used at output to obtain 750+ V DC for vacuum tube anode. My question is, why use a voltage doubler instead of using a appropriate transformer secondary voltage such as 600v RMS so that 750+ v can be obtained at output easily using normal bridge rectifier and filter without a voltage doubler. I am planning to build a transformer for tube RF linear amplifier and I wanted to use a transformer with secondary voltage of 600v RMS. I will use a bridge rectifier and capacitor filter to get 850 v DC and then use a regulator to get clean 720v DC. What should be constrain to make a 600v secondary transformer? Primary voltage is 220v. Below is the circuit.

• The circuit in your drawing is a bridge rectifier, not a voltage doubler. Each diode in the bridge uses two diodes in series, presumably for a higher voltage rating than a single diode. The main filter capacitor is also made up of three capacitors (C9, C10, and C11) also to get a higher voltage rating that may be available in a single capacitor. Jul 1 at 15:53

My question is, why use a voltage doubler instead of using a appropriate transformer secondary voltage such as 600v RMS so that 750+ v can be obtained at output easily using normal bridge rectifier and filter without a voltage doubler.

It's likely that the "old vintage tube power supply" design was based on an already available transformer and hence, using a couple of diodes and some capacitors to roughly double the voltage output is a sensible ploy.

Alternatively, put twice the secondary windings on and use a bridge rectifier. The advantage here is that each bridge diode only sees half the RMS current of the former design and, the secondary winding current rating is only half that of the former design.

Another reason might be that the original design needed a centre 0 volt point; that could be a reason why they chose the transformer that they did.

What should be constrain to make a 600v secondary transformer?

No real constraints that I can think of; just make sure that your diodes are rated for the correct working voltage plus another 30% for surges. If the DC output is 850 volts then the diodes need to be rated 30% more than twice this limit. That's 2200 volts. Of course you can put several in series to get the correct voltage rating.

• N.B. Because of variations among diodes, it is good practice to put large resistors and small capacitors across diodes in series to equalize voltage drop across them. One diode might have a back-impedance of 1 GΩ, and another 10 GΩ. Each would be fine by itself, but in series, one would have ~10% of the voltage across it, and the other ~90%. Jun 30 at 16:44
• @DrMoishePippik a point well made. Jun 30 at 20:27
• @DrMoishe Seeing a practical design for an extreme case is a nice-to-have. Here's how a physicist built a reversible 250 kV multiplier. Jun 30 at 23:51
• @DrMoishePippik yes i have done this. Sharing my circuit here. Its simulated from LTSPICE Jul 1 at 7:44
• @AshokDas if we are done here and you have your answer then please read this: What should I do when someone answers my question?. Of course if there's something that I haven't answered that appeared in your original question then please nudge me. Jul 1 at 9:13

Same power, double voltage will require same amount of copper in double the number of turns. However, truly halving the cross section area would also require a thinner lacquer isolation while requiring it to hold up to double the isolation voltage. Looking at typical transformer wire, I find (as an example) 1.7kV limiting voltage. Taken as peak voltage, this corresponds to 1200V RMS. However, the density of high voltage faults is certified with a test voltage of 500V, corresponding to just 360V RMS.

So creating a transformer with a certified output voltage of more than 360V RMS might require reverting to non-standard lacquer wire types, making for non-standard pricing and testing. Voltage doubling might be a cheaper solution here.