I've been using PSUD over the years but still some gaps, if someone can provide answers or anything I need to learn, here is where I'm at with the unregulated supply for a Papa Rusa...
1) For the DCR of the transformer I measured the DCR of my secondary to be 67 ohms. Should I also add in the DCR of the primary? Or is this setting mean the DCR of the secondary only.
2) In the simulation below I went with the default 2 ohms for the capacitors. These will all be DC link capacitors, what would be a common resistance to use here for film caps instead of 2 ohms, (which I assume are electros)?
3) I want to test how "fast" my power supply can recover from a 5ma upward transient draw. I want it to be "fast" so using as little capacitance and resistance as possible. This PS will be for a parafeed headphone amp (Papa Rusa link below) so I dont think the draw will be bouncing around much as its only driving headphones not speakers. So thats why I only tested a 5ma jump from 60ma to 65ma. As you can see it recovers pretty fast to the lower voltage and with no "bounce" or thrashing around just the downward dip, at the 2 second mark. I experimented with more capacitance but it was slower and the waveform hilly. It recovers in .4 seconds. Did I do such a transient simulation properly, or is my thinking way off here? The PS comes to full voltage in .6 seconds as well. I'm assuming the bounce at the beginning is just inrush because using solid state rectifiers, it seems to settle in fast though. Maybe the speed has to do with using a choke too instead of a resistor after C1, because chokes hold magnetic energy I would think thats better than a resistor right?
4) I want the ripple under 3mv, as you can see it achieved 2.8mv. This headphone amp uses a plate CCS so that will scrub even more ripple. So I think I'm happy with compromising ripple against speed.
5) Also note, the first choke is a real choke Hammond 14 H; 75 ma; 429 ohm model. But the second choke is a MOSFET gyrator set to simulate 30 H at 30 ohms. This headphone amp calls for 275V B+ @ 60ma so I think I nailed it with getting 280V. This transformer gives me exactly 236 volts from my home mains.
6) I took advantage of the highish DCR of the first choke to go with a smaller C1, much smaller than this and the PS seemed not stiff enough, I experimented with the middle capacitor but found that 1uf in the middle gave the best transient timing, the third cap I experimented with judging ripple against speed and settled upon 50uf which keeps me out of using electrolytics, the final 1 ohm and 3uf stage will be doubled for left/right channels.
This has been fun trying to make a power supply with as little resistance and capacitance as possible competing ripple against transient speed, it was good PSUD practice.
5ma upward transient 60 to 65ma at 2 seconds (looks good to me but I'm a newby):
No bouncing around at 2s just what seems like a nice smooth drop.
Ripple of 2.8 mv achieved:
The project:
Papa Rusa
https://wtfamps.com/papa-rusa-headphone-amplifier/
1) For the DCR of the transformer I measured the DCR of my secondary to be 67 ohms. Should I also add in the DCR of the primary? Or is this setting mean the DCR of the secondary only.
2) In the simulation below I went with the default 2 ohms for the capacitors. These will all be DC link capacitors, what would be a common resistance to use here for film caps instead of 2 ohms, (which I assume are electros)?
3) I want to test how "fast" my power supply can recover from a 5ma upward transient draw. I want it to be "fast" so using as little capacitance and resistance as possible. This PS will be for a parafeed headphone amp (Papa Rusa link below) so I dont think the draw will be bouncing around much as its only driving headphones not speakers. So thats why I only tested a 5ma jump from 60ma to 65ma. As you can see it recovers pretty fast to the lower voltage and with no "bounce" or thrashing around just the downward dip, at the 2 second mark. I experimented with more capacitance but it was slower and the waveform hilly. It recovers in .4 seconds. Did I do such a transient simulation properly, or is my thinking way off here? The PS comes to full voltage in .6 seconds as well. I'm assuming the bounce at the beginning is just inrush because using solid state rectifiers, it seems to settle in fast though. Maybe the speed has to do with using a choke too instead of a resistor after C1, because chokes hold magnetic energy I would think thats better than a resistor right?
4) I want the ripple under 3mv, as you can see it achieved 2.8mv. This headphone amp uses a plate CCS so that will scrub even more ripple. So I think I'm happy with compromising ripple against speed.
5) Also note, the first choke is a real choke Hammond 14 H; 75 ma; 429 ohm model. But the second choke is a MOSFET gyrator set to simulate 30 H at 30 ohms. This headphone amp calls for 275V B+ @ 60ma so I think I nailed it with getting 280V. This transformer gives me exactly 236 volts from my home mains.
6) I took advantage of the highish DCR of the first choke to go with a smaller C1, much smaller than this and the PS seemed not stiff enough, I experimented with the middle capacitor but found that 1uf in the middle gave the best transient timing, the third cap I experimented with judging ripple against speed and settled upon 50uf which keeps me out of using electrolytics, the final 1 ohm and 3uf stage will be doubled for left/right channels.
This has been fun trying to make a power supply with as little resistance and capacitance as possible competing ripple against transient speed, it was good PSUD practice.
5ma upward transient 60 to 65ma at 2 seconds (looks good to me but I'm a newby):
No bouncing around at 2s just what seems like a nice smooth drop.
Ripple of 2.8 mv achieved:
The project:
Papa Rusa
https://wtfamps.com/papa-rusa-headphone-amplifier/
Last edited:
I built the Papa Rusa exactly as the project describes it and I am very happy with the result. The CCS has a high impedance for AC signals so I would like to think the capacitors matter less to the sound here then in a typical amplifier. To me your analysis is sound but I am a novice at this as well. Your capacitor values are closer to what many vintage amps have used for much larger current loads.
Slightly off topic, I found the gain and bias of the 6s45P tubes to be all over the place tube to tube. The circuit auto biases as it should, but if you want exact gain matching I would add an extra volume trim pot or buy a larger set of tubes and match as best as you can.
Slightly off topic, I found the gain and bias of the 6s45P tubes to be all over the place tube to tube. The circuit auto biases as it should, but if you want exact gain matching I would add an extra volume trim pot or buy a larger set of tubes and match as best as you can.
Thanks,
I have a dozen or so 6S45P tubes and an eTracer so I will be able to curve match them, hopefully I'll get a few close ones! The article says 275V B+ and the voltage at the plate is 150V so that's a 125V drop but higher B+ is needed with a plate CCS I guess. Did you find the MOSFET to get very hot? Wondering if my heatsink is large enough.
I have a dozen or so 6S45P tubes and an eTracer so I will be able to curve match them, hopefully I'll get a few close ones! The article says 275V B+ and the voltage at the plate is 150V so that's a 125V drop but higher B+ is needed with a plate CCS I guess. Did you find the MOSFET to get very hot? Wondering if my heatsink is large enough.
I used some pretty large heatsinks and they get about as hot as the tubes, so you are right. I think this is well within the intended operating specs but if I had to do it again I would use larger heatsinks.