Shunt regs
I have never designed a shunt reg in my life if excluding zerers . To be honest it looks too good to be true that it has advantages . I would guess a slow transistor not especially wrong for a shunt if a capacitor backs it up . I have no idea and I am asking . I did do up a Chinese amp that had many of them . I can not claim it as my work . It did sound very good . It had a relay volume control using fixed resistors . The circuit resembled Naim .
Is there any need for speed ?
I have never designed a shunt reg in my life if excluding zerers . To be honest it looks too good to be true that it has advantages . I would guess a slow transistor not especially wrong for a shunt if a capacitor backs it up . I have no idea and I am asking . I did do up a Chinese amp that had many of them . I can not claim it as my work . It did sound very good . It had a relay volume control using fixed resistors . The circuit resembled Naim .
Is there any need for speed ?
the speed determines the maximum frequency the shunt reg can attenuate.
A very low frequency shunt circuit will be reasonably good at attenuating mains hum. A faster circuit will be good at attenuating mains buzz and LF noise.
A fast circuit will attenuate medium frequency noise, maybe even the noise below 20kHz.
A very fast shunt will attenuate high frequency noise, maybe to beyond 1MHz.
That's where the simulator shows when the output impedance is rising. When the attenuation of the HF noise is not so good.
If that paragraph is correct, then it follows that a very fast shunt is capable of attenuating more noise than any of the slower circuits.
A very low frequency shunt circuit will be reasonably good at attenuating mains hum. A faster circuit will be good at attenuating mains buzz and LF noise.
A fast circuit will attenuate medium frequency noise, maybe even the noise below 20kHz.
A very fast shunt will attenuate high frequency noise, maybe to beyond 1MHz.
That's where the simulator shows when the output impedance is rising. When the attenuation of the HF noise is not so good.
If that paragraph is correct, then it follows that a very fast shunt is capable of attenuating more noise than any of the slower circuits.
That's clear
Thanks Andrew . That makes perfect sense and never occurred to me . I guess shunt is a bit better than series as nothing is passing through an amplifying device directly between capacitor and amp lets say . My guess is two equally engineered series or shunt regulators would be nearly the same ? The difference being series is in fact a power amplifier swinging with the signal . The shunt also in a different way . Very hard to sum up I would say .
I would think shunt better if forced to use a slower transistor as it can be bolstered by the output cap ( e.g. 2N3055 ) . I suppose that holds for series also . The little power amp bobbing up and down of the series is not so appealing to my mind . LM317 for example can be turned into an amplifier .
Thanks Andrew . That makes perfect sense and never occurred to me . I guess shunt is a bit better than series as nothing is passing through an amplifying device directly between capacitor and amp lets say . My guess is two equally engineered series or shunt regulators would be nearly the same ? The difference being series is in fact a power amplifier swinging with the signal . The shunt also in a different way . Very hard to sum up I would say .
I would think shunt better if forced to use a slower transistor as it can be bolstered by the output cap ( e.g. 2N3055 ) . I suppose that holds for series also . The little power amp bobbing up and down of the series is not so appealing to my mind . LM317 for example can be turned into an amplifier .
That's very interesting Andrew . I used a double stage multiplier recently . It was only 22 mA . Yes now you say it I have seen it with JLH . I as an experiment used a near Victorian version to compare using high voltage and RC filters ( same final voltage ) . It slightly won and was about - 110 dB quiet . Very marginally the 1/f or Johnson noise of BD139/40 was higher than the high voltage version . That surprised me . The PSU had a double BD transistor as CCS to ensure safety and cap protection . Calculated surge current 190 A if not ( never happens in real life I know ) . That was because people do mad things which ends in smoke . This was to adapt a bought in product . We needed lower ripple noise . This is a LM317 pre-regulator . I have found 317/317 together can oscillate and RC is simple . I have a 47 V zener to catch the output when disconnected but still powered . The running output voltage is 22 V .
I think Douglas Adams said " The problem with making foolproof things is people underestimate the ingenuity of fools " . Something like that . Hence CCS .
Anyway that's me done with shunt regulators for now .
I think Douglas Adams said " The problem with making foolproof things is people underestimate the ingenuity of fools " . Something like that . Hence CCS .
Anyway that's me done with shunt regulators for now .
Noise filtering in parallel or regulation in parallel have a thing common (their being shunt) and two things uncommon. Those are a reg has error amplification and voltage reference. The reg usually achieves lower and flatter output impedance. Combinations of filtering and regulating aren't mutually exclusive though.
Talon
Hi Salas . I just accept Shunt Regulators . However the more I think about them the more confused I get .
The Chinese amp was Incatech Talon . Shame they didn't go ahead . I would have bought one myself . I know the designer but won't say who it is . Fantastic power supply . I have the drawings somewhere
Inca Designs Talon vs Incatech Claymore - pink fish media
It was a work of art and designed by a power supply obsessive . The most beautifully textured sound . Like a natural hi end successor to A+R A60 . The one they forgot to make .
Hi Salas . I just accept Shunt Regulators . However the more I think about them the more confused I get .
The Chinese amp was Incatech Talon . Shame they didn't go ahead . I would have bought one myself . I know the designer but won't say who it is . Fantastic power supply . I have the drawings somewhere
Inca Designs Talon vs Incatech Claymore - pink fish media
It was a work of art and designed by a power supply obsessive . The most beautifully textured sound . Like a natural hi end successor to A+R A60 . The one they forgot to make .
Last edited:
Hi Andrew . I have difficultly with that also . I tried the JLH idea and found Early effect a problem if not using the small bleed resistor to ground to ensure the transistors are not in saturation . I used a divider on mine . Some use a diode forward biased CB to ensure Early effect doesn't happen . Also a back EMF diode to CE and BC reverse biased ( 1N4007 my do everythinng diode ) . They cost 2 pence so why not ?
Do you think MOS FET better than Bipolar in Shunt ? I am trying to shunt 100 MHz . The idea only came to me after your comment . I know you are methodical so trust your advice .
Thanks .
Do you think MOS FET better than Bipolar in Shunt ? I am trying to shunt 100 MHz . The idea only came to me after your comment . I know you are methodical so trust your advice .
Thanks .
Me too and I am loosing the battle . I think a ceramic cap is all I can do ( 1 nF COG )
Nigel
Don't know about the extra inductance on those wires. Never done it like that, can't comment without experiment. If talking SSHV2 there is 5.5MHz open loop zero gain point extension. Loss of phase margin or interference can happen. It will take trial and measurement so you see if still safe with your total configuration. Sensing compensates to the remote node, does not shorten the local force loop area.