Most of the guys doing this found they liked the sound of mosfets in the multiplier over bipolar transistors.
Its prefiltering, so it can be good, why not. Don't know about the extra volts needed though if trafo is tight. CMC coils can be good too. I was thinking of a full power BJT cascode CCS instead of the mini only. That would take psrr performance higher for far higher frequencies. I was thinking of a depletion cascode too but then again not much internal feedback, or of an enhancement cascode with 10V Vgs + Vgs that is excellent, ala G. Pimm, could be done with leds, and will go HV too, but then again too much voltage loss. Or a cascode mosfet gyrator? What do you think? I guess will try and listen for a V1.2 at a point.
Yes, definitely just pre-filter. I would stay away from the lm317, I know other people use it, but it's just my preference.
For high frequency psrr everywhere I looked outside the audio field, guess what they use. Ferrite inductors, and common mode chokes. Simple, cheap, and efficient.
Then, two gyrators (or cap multipliers) in series, if one has enough juice in the transformer. There should be enough psrr even for the very obsessed. The noise getting in will probably be only that generated by the shunt reg itself, or modulated by the powered circuit.
For high frequency psrr everywhere I looked outside the audio field, guess what they use. Ferrite inductors, and common mode chokes. Simple, cheap, and efficient.
Then, two gyrators (or cap multipliers) in series, if one has enough juice in the transformer. There should be enough psrr even for the very obsessed. The noise getting in will probably be only that generated by the shunt reg itself, or modulated by the powered circuit.
Talk about radio interference and its effect on audio stuff. here I quote the view of someone, from a presentation he made:
"The AudioFool Viewpoint
• The Myth: “We can hear stuff above 20 kHz”
• Reality: Some distortion mechanisms DO
produce audible artifacts from ultrasonic
signals, but we hear the problems, not the
signals!
• Intermodulation distortion (40kHz – 30kHz = 10 kHz)
• Slew rate limiting within electronics of
ultrasonic output of a mic (or of square
waves from a test generator)"
It's an entertaining document
http://audiosystemsgroup.com/Ferrites-Ham.pdf
😀
"The AudioFool Viewpoint
• The Myth: “We can hear stuff above 20 kHz”
• Reality: Some distortion mechanisms DO
produce audible artifacts from ultrasonic
signals, but we hear the problems, not the
signals!
• Intermodulation distortion (40kHz – 30kHz = 10 kHz)
• Slew rate limiting within electronics of
ultrasonic output of a mic (or of square
waves from a test generator)"
It's an entertaining document
http://audiosystemsgroup.com/Ferrites-Ham.pdf
😀
"Just spent 3 hours listening to the FET cap multiplier on speakers. I even managed to listen to a whole Coldplay album in one sitting, its that good. System is CDS3, 62 with 2 X Superregs and 2 X R Core and Mr Tibbs' earthing mod. 62 is standard. 2 X NAP 135's and PMC FB1's.
It is in my view a big improvement on the darlington. It is much easier to hear down into the mix, and the tonal balance just seems right all the time for once. Tracks that didn't appeal before, do now. I think the bass and lower mid now operate independantly and without modulating the finer elements of the top end.
An example. On Pink Floyd's Money, during the saxaphone riff I've always been aware of some very deep bass, more felt than heard. I knew it stopped during the solo but only now can I pin point the exact moment. Piano music is improved, with the whole range of the keyboard exhibiting the same pianos characteristics. It sounds like a coherant instrument. There is a deeper sonic blackness behind most tracks now and fades appear to continue longer. If there is a downside it is a hint of softness in the balance of some bass tracks. It is not a slowness and it may be correct now and just different to what I had been used to.
For me the switch from a 1Mhz Darlington to this IRF 520 FET was highly beneficial. About the same improvement as going from the (LM317) pre reg to the darlington Vbe." John.Luckins pinkfishmedia
It is in my view a big improvement on the darlington. It is much easier to hear down into the mix, and the tonal balance just seems right all the time for once. Tracks that didn't appeal before, do now. I think the bass and lower mid now operate independantly and without modulating the finer elements of the top end.
An example. On Pink Floyd's Money, during the saxaphone riff I've always been aware of some very deep bass, more felt than heard. I knew it stopped during the solo but only now can I pin point the exact moment. Piano music is improved, with the whole range of the keyboard exhibiting the same pianos characteristics. It sounds like a coherant instrument. There is a deeper sonic blackness behind most tracks now and fades appear to continue longer. If there is a downside it is a hint of softness in the balance of some bass tracks. It is not a slowness and it may be correct now and just different to what I had been used to.
For me the switch from a 1Mhz Darlington to this IRF 520 FET was highly beneficial. About the same improvement as going from the (LM317) pre reg to the darlington Vbe." John.Luckins pinkfishmedia
Last edited:
I don't know, maybe I will make the low mc with new enhanced shunt and then think I wanna listen to it on a new speaker too. Horrors.😀who knows what will come out now?
I built some fet cap multipliers before, it's an easy circuit conceptually. The only reason why I showed the bjt was because it was handy in ltspice (didn't need to search for it). In reality I'd use a mosfet for the pass element because, among other reasons, it's so much easier to get drive when high current is needed.
http://i27.photobucket.com/albums/c156/princesnuggles/FETCAPMULTIPLIER.jpg
Must use polymer caps to get the ultimate performance (except for the 15000uF cap of course).
Must use polymer caps to get the ultimate performance (except for the 15000uF cap of course).
Last edited:
Thanks. Hmm I think I have seen that before in the past somewhere. Sure it's an option for not to overcook the main ccs for psrr. Pre filtering keeps things simpler.
"Carl, like you, I have gone through almost all of the versions of this gyrator circuit, and arrive at the same conclusion: stick with the FET.
I have now directly compared BJT Darlington (547C/D44H11) and a FET (IRF 520) gyrator pre-reg in a number of circuits. Both versions were built with carefully constructed layout with minimum trace length between the filter components and approx. 5mm between last filter resistor/cap combo and relevant BJT/FET leg. Each time I have compared the two I have been impressed by how open, wide-band, uncolored, and precise the FET version performed. There is also a tremendous solidity and power to the presentation that is unlike the BJT Darlington versions. I do not know if the sound will get more free and mellower with time, but if you like rock music, the FET version is definitely for you.
I also designed one blind-fold A/B test to compare the BJT and FET control element within the same circuit (AD825-based ADC buffer) with two other audiophiles. They both preferred the FET version. The most striking feature of the FET-based version in this circuit was better LF extension, an aspect which is by no means trivial in a circuit packed with metal foil resistors and polyprop coupling caps. By comparison, the BJT pre-reg did less well at low frequencies, but had a deeper sound-stage for some reason, in part related to the sharper way it reproduced mid and higher frequencies. Everyone agreed that this was however an artificial effect, and that the FET version sounded "more linear" and "more rhythmically and tonally accurate."
It appears that the constant LF impedance of the FET version is relevant from a sonic and a from a power distribution quality point of view. By quality, I mean both the magnitude of the impedance, plus how well it is maintained over a large frequency range. Even if their very high voltage threholds severely limit attempts for low drop-out designs,the linearity of the FET output impedance and its dynamic performance make it very attractive as a control element in a gyrator pre-reg.
From this moment on, I am devoting my attention only to FET-based gyrators, even going through the hassle of replacing transformers in equipment where the transformer secondary voltage is too low to accomodate the higher drop-out voltage of the FET version. Catalogs are being consulted to source other varieties of FET to test against the IRF520, and plans are being made to implement TO-92-style FET gyrators in low-power regulation systems (including a flea or two ). It will be very interesting to see what sort of characteristics and performance the gyrator design results with different FETs. I hope to have some results to report in a month or two." Slawney pinkfishmedia
I have now directly compared BJT Darlington (547C/D44H11) and a FET (IRF 520) gyrator pre-reg in a number of circuits. Both versions were built with carefully constructed layout with minimum trace length between the filter components and approx. 5mm between last filter resistor/cap combo and relevant BJT/FET leg. Each time I have compared the two I have been impressed by how open, wide-band, uncolored, and precise the FET version performed. There is also a tremendous solidity and power to the presentation that is unlike the BJT Darlington versions. I do not know if the sound will get more free and mellower with time, but if you like rock music, the FET version is definitely for you.
I also designed one blind-fold A/B test to compare the BJT and FET control element within the same circuit (AD825-based ADC buffer) with two other audiophiles. They both preferred the FET version. The most striking feature of the FET-based version in this circuit was better LF extension, an aspect which is by no means trivial in a circuit packed with metal foil resistors and polyprop coupling caps. By comparison, the BJT pre-reg did less well at low frequencies, but had a deeper sound-stage for some reason, in part related to the sharper way it reproduced mid and higher frequencies. Everyone agreed that this was however an artificial effect, and that the FET version sounded "more linear" and "more rhythmically and tonally accurate."
It appears that the constant LF impedance of the FET version is relevant from a sonic and a from a power distribution quality point of view. By quality, I mean both the magnitude of the impedance, plus how well it is maintained over a large frequency range. Even if their very high voltage threholds severely limit attempts for low drop-out designs,the linearity of the FET output impedance and its dynamic performance make it very attractive as a control element in a gyrator pre-reg.
From this moment on, I am devoting my attention only to FET-based gyrators, even going through the hassle of replacing transformers in equipment where the transformer secondary voltage is too low to accomodate the higher drop-out voltage of the FET version. Catalogs are being consulted to source other varieties of FET to test against the IRF520, and plans are being made to implement TO-92-style FET gyrators in low-power regulation systems (including a flea or two ). It will be very interesting to see what sort of characteristics and performance the gyrator design results with different FETs. I hope to have some results to report in a month or two." Slawney pinkfishmedia
Maybe that IRF520 cct before v1.1 is something to listen too. Must be droping 4 Volt? I have some IRF820s. Low crss.
- Status
- Not open for further replies.