A little background (skip if this is boring): I wanted to make a power supply for a tube circuit, using two ordinary transformers back to back:
230V:25V (transformer #1) -> 25V:230V (transformer #2)
Works nicely, it's often done, my version is here:
http://www.oldcrows.net/~jhaible/opto_compressor/jh_schematics_4_psu.jpg
Nothing special - the downside is that I only get ca. 230V instead of the 250V that were called for in the circuit I wanted to clone. Also, 250V seem to be quite common in the tube world.
So my idea was to connect transformer #2 as an auto-transformer (not shown in the drawing, but you know what I mean), connecting primary and secondary of #2 in series, thus adding the 25V to the 230V.
The problem I saw is the isolation voltage. As in that specific application, I wanted to use the second secondary of transformer #1 for other parts of the circuit (low voltage, near GND level), so with #2 not isolating anymore (auto-transformer ...), I would need the whole isolation voltage between first secondary and second secondary of transformer #1. So I emailed Talema to get the data for that specific isolation voltage (I'll come back to that below), and of course the isolation is not good enough. It's just two wires wound in bifilar way - no real isolation. They test it with 500V, which is next to nothing compared to 4200V test voltage between primary and secondary. (Remember the test voltage must be many times higher than the voltage in your application !!)
So the auto-transformer idea does not work. (It will work if you don't use two separate secondaries on transformer #1. For instance, if you use a separate, 3rd transformer for tube heaters etc.)
There is another method that will work, however: Use two rectifiers, one for the intermediate 25V, and one for the second-secondary 230V, and then stack the two pulsing DC voltages (where the rectified 25V must be connected near GND, not the other way round, for obvious reasons).
Simulations have shown that you don't even need two capacitors for stacking the two rectified voltages: Just stack the raw rectifier outputs and run the combined voltage into your typical 450V-rated capacitor. I have *not* tested this yet, but simulation is promising, and isolation voltage apparently is good enough for this.
I don't think I'm th efirst to do this, but for me it was brand new, so I thought I'd share the idea.
Ok, back to talema - I'll make a follow-up posting for this.
JH.
230V:25V (transformer #1) -> 25V:230V (transformer #2)
Works nicely, it's often done, my version is here:
http://www.oldcrows.net/~jhaible/opto_compressor/jh_schematics_4_psu.jpg
Nothing special - the downside is that I only get ca. 230V instead of the 250V that were called for in the circuit I wanted to clone. Also, 250V seem to be quite common in the tube world.
So my idea was to connect transformer #2 as an auto-transformer (not shown in the drawing, but you know what I mean), connecting primary and secondary of #2 in series, thus adding the 25V to the 230V.
The problem I saw is the isolation voltage. As in that specific application, I wanted to use the second secondary of transformer #1 for other parts of the circuit (low voltage, near GND level), so with #2 not isolating anymore (auto-transformer ...), I would need the whole isolation voltage between first secondary and second secondary of transformer #1. So I emailed Talema to get the data for that specific isolation voltage (I'll come back to that below), and of course the isolation is not good enough. It's just two wires wound in bifilar way - no real isolation. They test it with 500V, which is next to nothing compared to 4200V test voltage between primary and secondary. (Remember the test voltage must be many times higher than the voltage in your application !!)
So the auto-transformer idea does not work. (It will work if you don't use two separate secondaries on transformer #1. For instance, if you use a separate, 3rd transformer for tube heaters etc.)
There is another method that will work, however: Use two rectifiers, one for the intermediate 25V, and one for the second-secondary 230V, and then stack the two pulsing DC voltages (where the rectified 25V must be connected near GND, not the other way round, for obvious reasons).
Simulations have shown that you don't even need two capacitors for stacking the two rectified voltages: Just stack the raw rectifier outputs and run the combined voltage into your typical 450V-rated capacitor. I have *not* tested this yet, but simulation is promising, and isolation voltage apparently is good enough for this.
I don't think I'm th efirst to do this, but for me it was brand new, so I thought I'd share the idea.
Ok, back to talema - I'll make a follow-up posting for this.
JH.