Hello everyone, is there any reason I couldn’t have a capacitor multiplier with multiple pass elements?
I’m in the early stages of designing a power supply for an amp i’m working on where the B+ can be adjusted. I was designing one with one B+ output for both channels, but now i’m wondering if I can get better channel isolation by "slapping on" another pass element running off the same driver, since the pass elements have high impedance inputs.
I could run two separate circuits off the main filter cap but then I would have two voltage pots which I prefer to avoid.
I’m in the early stages of designing a power supply for an amp i’m working on where the B+ can be adjusted. I was designing one with one B+ output for both channels, but now i’m wondering if I can get better channel isolation by "slapping on" another pass element running off the same driver, since the pass elements have high impedance inputs.
I could run two separate circuits off the main filter cap but then I would have two voltage pots which I prefer to avoid.
Why? Why don't yopu run both followers from the same adjusted voltage? They should be the same anyway, no?
Jan
Jan
im trying to lower the effect of dynamics/impulse of one channel affecting the B+ on the other channel
Yes I get that. You think that therefor the two B+'s must be different?
The way you draw it you get two indepent source followers fed from the same ref voltage.
Isn't that what you want?
Jan
The way you draw it you get two indepent source followers fed from the same ref voltage.
Isn't that what you want?
Jan
Yep - That’s exactly what I want - I’m just asking here if that’s an « acceptable thing to do » since I haven’t seen it in many other schematics before. I figure there must be a reason I might be ignorant to at the moment. A lot of the time I try something "weird" I usually arrive to "oooh that’s why they don’t do that" after having spent a week playing with something on the bench. I’m also hoping some people can point out some potential points of problems or instabilities that I should be looking out for if I go ahead with the design. To me this is important so that when i’m designing the PCB, I can design it for ease of parts-swapping/alterations in areas that are questionable so that’s that they can be refined for the next version if the PCB.
One potential issue is thermal management.. I guess it’s not really an issue if i’m being reasonable with how much i’m dropping the B+, but i’m still a bit worried about the ballast resistors on the capacitors dissipating a lot of heat, those resistor values are somewhat arbitrary for now, i’ll have to calculate them based on what voltage I find myself setting the supply at most often.
I’m also considering using IGBTs for the pass elements instead of MOSFETS for the lower output impedance.
I am interested in trying out this kind of circuit for tubes. The idea to get better stereo separation using distinct regulator branches is interesting. I guess I don't really understand the 'multiplication' effect on capacitance, but it seems to me this is just 'apparent' capacitance that reduces the ripple but it does not provide any real energy storage. In other words would you still need to add large caps on each B+ branch?
How is IGBT different from mosfet to provide lower output impedance?
How is IGBT different from mosfet to provide lower output impedance?
MOSFETs inherently have an higher output impedance than BJTs or IGBTs.
Solid state amps with a MOSFET output stage sometimes regarded as “softer” sounding than BJT output stages.
From what I understand, IGBTs are kind of like the best of both worlds of combining the high impedance voltage control of MOSFETs and the low output impedance of BJTs… IGBTs are also commonly available in high voltage and current handling types… With the caveat being that they’re slow devices not really suited for use beyond 20Khz, which is perfectly fine for power supplies.
I used an IGBT because I wanted a pass device with a high impedance so I could use a potentiometer to control the voltage without burning the track. A MOSFET would have worked just fine, but I figured an IGBT might be give me a stiffer supply.
C3/7/12 are the main filter caps. These should still be big to give the pass elements a decent reserve.
the two columns of caps to the right of it make up the RC filter part of the “capacitor multiplier” which acts as part of the driver circuit for the pass elements. Q3 is there for varying the output voltage… but this needs to be done with care since it could cause the pass elements to dissipate/waste a lot of energy as heat. It’s mainly there to trim out anything I can’t achieve by playing around with different transformers/taps
Solid state amps with a MOSFET output stage sometimes regarded as “softer” sounding than BJT output stages.
From what I understand, IGBTs are kind of like the best of both worlds of combining the high impedance voltage control of MOSFETs and the low output impedance of BJTs… IGBTs are also commonly available in high voltage and current handling types… With the caveat being that they’re slow devices not really suited for use beyond 20Khz, which is perfectly fine for power supplies.
I used an IGBT because I wanted a pass device with a high impedance so I could use a potentiometer to control the voltage without burning the track. A MOSFET would have worked just fine, but I figured an IGBT might be give me a stiffer supply.
C3/7/12 are the main filter caps. These should still be big to give the pass elements a decent reserve.
the two columns of caps to the right of it make up the RC filter part of the “capacitor multiplier” which acts as part of the driver circuit for the pass elements. Q3 is there for varying the output voltage… but this needs to be done with care since it could cause the pass elements to dissipate/waste a lot of energy as heat. It’s mainly there to trim out anything I can’t achieve by playing around with different transformers/taps
Is this output impedance you speak of related to Rds ON? I see that the apt8100m you chose has 1.5ohm, there are SiC devices out there that have 30-50mohm RdsON. The power dissipation would be lower.