Is it possible to design a power supply for a tube amplifier that will only filter frequencies 5khz and up? Would it necessarily require the same filtration as a full range amp?
In other words, since the filtration of a full range amp has to filter out 60hz, 120hz and its fundamentals, and if the high frequency amp only has 50khz and up passing though it, would it need a "standard" p/s?
In other words, since the filtration of a full range amp has to filter out 60hz, 120hz and its fundamentals, and if the high frequency amp only has 50khz and up passing though it, would it need a "standard" p/s?
The hum will come through on the output, and also cause IMD. Excessive LF power supply ripple
will also overheat the power supply capacitors if they are too small.
will also overheat the power supply capacitors if they are too small.
Last edited:
What if there was a first order filter to prevent say 5khz and below on the amp itself to prevent the hum and IMD from passing to the speakers?
When you say small, are you referring to their capacitance or their voltage rating? Would increasing either or both help prevent overheating?
Thanks.
Very good question, actually. IMHO this could work out if the amplifier's feedback is designed such that frequency response is pushed down close to unity at 120 Hz, aiding PSRR down there. It's probably not the kind of difference you might expect but up to 20 dB more ripple tolerance should be doable.
But as rayma quite correctly mentions, if amplifier gain varies with supply voltage, you will get amplitude modulation of the signal to be amplified, i.e. IMD. Using feedback in whatever form to pull gain down there down near unity would address this. A suitable choice of coupling capacitors would reduce the amount of hum reaching the tweeter but (IMHO) do nothing about IMD.
But as rayma quite correctly mentions, if amplifier gain varies with supply voltage, you will get amplitude modulation of the signal to be amplified, i.e. IMD. Using feedback in whatever form to pull gain down there down near unity would address this. A suitable choice of coupling capacitors would reduce the amount of hum reaching the tweeter but (IMHO) do nothing about IMD.
Last edited:
Thank you for your response. I am not knowledgable enough about these things to have even a humble opinion, but I would like to think that there is something workable in the idea.
And this is why I ask: Presently I have a full range amplifier which I built and I would like to add a simple "tweeter amplifier" to it, separate the frequencies with a small 1st order crossover allowing the tweeter amp to handle everything over 5khz. I do not want to biamp; I want to integrate as much as possible the two amps, at least in one chassis. This would also alleviate the use of a crossover in the speakers.
The two amps would/could(?) share tubes-- though the output tubes obviously would be different--but the high frequency amp would have have it own separate p/s -- and therefore the question, thinking that it would be physically smaller than the full range p/s.
The limitations are these: if I wanted to run both amps on the one power supply, with the current requirements for both amps, the caps and most importantly the chokes in the power supply would be too big and take up more real estate than I had to offer. Also I would like to stay away from any feedback design that would limit dynamics.
Is any of this possible?
And this is why I ask: Presently I have a full range amplifier which I built and I would like to add a simple "tweeter amplifier" to it, separate the frequencies with a small 1st order crossover allowing the tweeter amp to handle everything over 5khz. I do not want to biamp; I want to integrate as much as possible the two amps, at least in one chassis. This would also alleviate the use of a crossover in the speakers.
The two amps would/could(?) share tubes-- though the output tubes obviously would be different--but the high frequency amp would have have it own separate p/s -- and therefore the question, thinking that it would be physically smaller than the full range p/s.
The limitations are these: if I wanted to run both amps on the one power supply, with the current requirements for both amps, the caps and most importantly the chokes in the power supply would be too big and take up more real estate than I had to offer. Also I would like to stay away from any feedback design that would limit dynamics.
Is any of this possible?
The power supply has to provide a (close to) constant voltage as the mains cycles, this is the same whether the amp is outputing 100Hz, 10kHz or 1MHz. It stores energy for during the time the mains voltage is zero-crossing, so the capacitors are sized dependent on the (average) power drawn and the mains frequency.