Thank you... just what I was needing
I always felt the stylus tip to be one of the major factors regarding vinyl playback. I used to swap carts just to analyse the différences I got with different tips... Remenber SP12 ?
At the moment i have a cartridge with an original Shibata needle ( Ortofon MC 30 super export modified by a local cartridge maker ) and a Lyra Titan i with an Ogura double polished professional needle.
Maybe i will let refurbish my Audio Technica with a micro ridge.
I do not think that grinding a needle any sharper then the Shibata brings any benefit in sound but the double polish of the Ogura and the long ridge of the Micro ridge may enhance life span. Anyway, it is mostly the suspension that gives up, long before the needle is worn. 3000 hors life span is the norm with a good, natural diamond.
Maybe i will let refurbish my Audio Technica with a micro ridge.
I do not think that grinding a needle any sharper then the Shibata brings any benefit in sound but the double polish of the Ogura and the long ridge of the Micro ridge may enhance life span. Anyway, it is mostly the suspension that gives up, long before the needle is worn. 3000 hors life span is the norm with a good, natural diamond.
My early Kondo Audionote needs work! Now that Kondo is no longer with us, who is the best guy to do a cantilever + tip job? I had a new tip fitted eighteen months ago and even I can hear that it is not right........it sounded better before that work!
I also HAD an Ortofon MC20 70th anniversary (with ceramic body) which went for similar work (to another cartridge repairer) and it has not been returned. They say it was posted, but uninsured and no tracking details!
I also HAD an Ortofon MC20 70th anniversary (with ceramic body) which went for similar work (to another cartridge repairer) and it has not been returned. They say it was posted, but uninsured and no tracking details!
This guy is a master :
| Retipping Service | Wir reparieren Ihr hochwertiges MC | MM System
The prices are moderate.
| Retipping Service | Wir reparieren Ihr hochwertiges MC | MM System
The prices are moderate.
Hi,
the Rmes in #9073 are for measuring purposes only. The Rser is the actual internal resistance of one of the voltage sources, here V2.
These solved and anything else seems better than with classic PSs.
Medicine-grade SMPS reduce the leakage currents vastly.
jauu
Calvin
the Rmes in #9073 are for measuring purposes only. The Rser is the actual internal resistance of one of the voltage sources, here V2.
....but why is there kind of ´sweet spot´ of most linear Zout and a minimum off of 0R Rser? Wouldn´t one expect the lowest Zout with the lowest internal voltage source impedance?
That and the PE-Gnd issues are imho the big issues with certain SMPS.
These solved and anything else seems better than with classic PSs.
Medicine-grade SMPS reduce the leakage currents vastly.
jauu
Calvin
That is an interesting observation. This could be an unintended "feature" of the Kmultiplier. The question is, does it improve anything? Play with different models and difference reference resistance values and see what happens.
Just guessing; At this very low Vce the voltage margin available for 'control' gets very small (Vce - Vbe (as Vce is also Vpsu)), somehow this is influencing behavior of the whole thing. The more current you ask the higher Vbe, the Hfe lessening, and some balance point will be in there (I guess
). Just my opinion, I'm a simulant (user of simulators )Medicine-grade SMPS's, that's a good, but somewhat expensive, thought.
Frans.
Last edited:
Some good questions to ask are:
1: Does driver Ib increase with loading below 1Hz as expected or does it decrease?
2: Does Vbe increase with loading below 1Hz as expected or does it decrease?
3: Either reference resistor current must be 180 degrees off or Vbe must be 180 degrees off. Include phase in a separate pane on plots, as this will help sort out polarity questions. It could be that resistance actually becomes negative at some point, but again without a phase plot we will never know. In Bode plots (polar coordinates) there are no negative values, instead there is phase.
1: Does driver Ib increase with loading below 1Hz as expected or does it decrease?
2: Does Vbe increase with loading below 1Hz as expected or does it decrease?
3: Either reference resistor current must be 180 degrees off or Vbe must be 180 degrees off. Include phase in a separate pane on plots, as this will help sort out polarity questions. It could be that resistance actually becomes negative at some point, but again without a phase plot we will never know. In Bode plots (polar coordinates) there are no negative values, instead there is phase.
Keantoken, Calvin, and everyone else,
OTH I tried and tried and ... but did not succeed to improve the K-multiplier! (I just want to make that clear). But then came the question to me, should we use this to feed an audio line driver (or pre amp). The happy thing here is that this does not deal with very high voltages and powers, so it is allowable to drop a few more volts. That made me think that a design (like the K-multiplier) with less components could have the same performance (except for the low voltage drop). This is why I designed the 'simple' version.
Joachim will build both versions and test, taken FFT's and more (I'm still curiously awaiting the results)
OTH I do not think that such a simple design is good enough for powering an audio line driver or pre amp. Most probably there are 'standard'-regulators that may be more fitting and still have a low component count. As Hesener showed you can improve the simple design(s) that I did show easily, if you do not mind the DC ground reference, but then you add complexity (components).
Then there is still the open question, what is the best solution (powering an audio line driver, pre amp, or RIAA)? Any way I (for one) will be working to answer that, currently there are ideas (these (at the least the one's by me) will be developed, tested and then published).
OTH I tried and tried and ... but did not succeed to improve the K-multiplier! (I just want to make that clear). But then came the question to me, should we use this to feed an audio line driver (or pre amp). The happy thing here is that this does not deal with very high voltages and powers, so it is allowable to drop a few more volts. That made me think that a design (like the K-multiplier) with less components could have the same performance (except for the low voltage drop). This is why I designed the 'simple' version.
Joachim will build both versions and test, taken FFT's and more
(I'm still curiously awaiting the results)OTH I do not think that such a simple design is good enough for powering an audio line driver or pre amp. Most probably there are 'standard'-regulators that may be more fitting and still have a low component count. As Hesener showed you can improve the simple design(s) that I did show easily, if you do not mind the DC ground reference, but then you add complexity (components).
Then there is still the open question, what is the best solution (powering an audio line driver, pre amp, or RIAA)? Any way I (for one) will be working to answer that, currently there are ideas (these (at the least the one's by me) will be developed, tested and then published).
Last edited:
Some papers about designing low noise switching power supply from Linear Technology, one of them was written by Jim Williams.
http://cds.linear.com/docs/en/design-note/dn316f.pdf
http://cds.linear.com/docs/en/application-note/an70.pdf
http://cds.linear.com/docs/en/design-note/dn316f.pdf
http://cds.linear.com/docs/en/application-note/an70.pdf
Sorry, i ran into timing problems so middle of the week is earliest i can test various cap multipliers.
I have build the Frans Simple but it got hot and was drawing 1A with 220 Ohm load resistor. That was the one with 68kOhm and 4.7uF foil.
The PSU was my lab supply without input cap to the multiplier and 10uF Panasonic FM shunt at the output.
Frans do you have any idea how i could stop the oszillation ?
I have build the Frans Simple but it got hot and was drawing 1A with 220 Ohm load resistor. That was the one with 68kOhm and 4.7uF foil.
The PSU was my lab supply without input cap to the multiplier and 10uF Panasonic FM shunt at the output.
Frans do you have any idea how i could stop the oszillation ?
Your're right. The Kmultiplier is more of a filter. It is designed to have high RF PSRR while being basically inert.
I have a friend who has compared the Kmultiplier favorably to the LM337/317, in many ways at least. Look at this 4-part series on chipregs which may help to understand why:
Using 3-pin regulators off-piste: part 3
Any unconditionally stable high-feedback regulator will have significant output inductance which may resonate with low-ESR lytics at the output. May people obsess over low-Zout regulators but ignore output inductance. The result is a high-Q resonator in the audio band, and maybe rampant fiddling to put that peak somewhere benign. However if you make output resistance roughly equivalent to the lytics in use, this problem fixes itself because there is no need for high feedback and high output inductance. This is what I have done with the Kmultipliers. I did not expect to come up with something revolutionary, and this freedom allowed me to apply the principle of "first do no harm".
If you need something revolutionary, there are plenty of other regulators to choose from like the Salas regs or the Ikoflexer variants. However many high-feedback regs like this cannot tolerate local bypass, which is really the only way to reduce RF rail impedance, so we are back to square one.
I think in the end it takes a lot of skill to apply even a Kmultiplier in the best possible way. High-feedback regulators are even harder to implement optimally and are less forgiving. With a Kmultiplier you can throw a bunch of lytics at it and you're less likely to have to fiddle.
So, what do you guys think? What are the actual advantages to using something more than a Kmultiplier, in terms of practical effects, and how much extra complexity is it worth?
I have a friend who has compared the Kmultiplier favorably to the LM337/317, in many ways at least. Look at this 4-part series on chipregs which may help to understand why:
Using 3-pin regulators off-piste: part 3
Any unconditionally stable high-feedback regulator will have significant output inductance which may resonate with low-ESR lytics at the output. May people obsess over low-Zout regulators but ignore output inductance. The result is a high-Q resonator in the audio band, and maybe rampant fiddling to put that peak somewhere benign. However if you make output resistance roughly equivalent to the lytics in use, this problem fixes itself because there is no need for high feedback and high output inductance. This is what I have done with the Kmultipliers. I did not expect to come up with something revolutionary, and this freedom allowed me to apply the principle of "first do no harm".
If you need something revolutionary, there are plenty of other regulators to choose from like the Salas regs or the Ikoflexer variants. However many high-feedback regs like this cannot tolerate local bypass, which is really the only way to reduce RF rail impedance, so we are back to square one.
I think in the end it takes a lot of skill to apply even a Kmultiplier in the best possible way. High-feedback regulators are even harder to implement optimally and are less forgiving. With a Kmultiplier you can throw a bunch of lytics at it and you're less likely to have to fiddle.
So, what do you guys think? What are the actual advantages to using something more than a Kmultiplier, in terms of practical effects, and how much extra complexity is it worth?
You can add a small (2...5 ohm go as small as possible) series resistor to the 'Panasonic FM' or/and the driver transistor may be compensated by a bc capacitor of 1..5nF (but that will impact HF performance). One other thing I can think off is to add a base stopper resistor.
Here is the pass transistor :
http://www.alphacron.de/download/hardware/BC161.pdf
It may be too slow, it has also has a rather high Cob.
I could use the Zetex that Calvin uses in his simulation.
http://www.alphacron.de/download/hardware/BC161.pdf
It may be too slow, it has also has a rather high Cob.
I could use the Zetex that Calvin uses in his simulation.