A brief bit about power supply design.
Let us be clear about the difference between a power amplifier power supply requirement and other lower level gear.
A high end power amplifier typically will have very large power supply filter capacitor banks. These will reduce ripple and in theory allow large peak currents. In practice if coupled with an efficient transformer they will place very large peak current demands on the AC power line. That is because low ripple also means less available time at the AC cycle peak to charge. This higher demand current makes issues such as AC line resistance much more important. A typical very good AC line and power cord will have .2 ohms of resistance. If you consider that 100mv of ripple on a 50 volt supply is 500 to one, you begin to understand why .2 ohms can be a big deal. (The inductance in the transformer reduces this 500 x peak)
The front end of a power amplifier has very similar needs for stable voltages as do the other small signal pieces of gear.
So the first issue for small signal gear is what happens to the internal power supply when the power amplifier is clipping the AC power line voltage peaks?
Almost any regulator will solve the voltage issue, but narrow current draw pulses also make noise!
Some will place a capacitor across the incoming AC line to reduce some of the line noise issues.
It is obvious that some of the noise makes it past the power supply transformer. Shields will of course reduce noise but never eliminate it. As most who have encountered an internal transformer shield know there is an issue of how you ground it and does the safety ground reduce or increase the noise.
After the transformer are of course rectifier diodes. The wise high end power supply designer may use special diodes to reduce switching noise, some may even place small resistors here to smooth and stretch currents, some even go so far as to use parallel diode resistor combinations.
The filter capacitors for those who have looked have internal resistance and inductance. This limits their effective filtering action. Most will place smaller capacitors across larger ones to get greater bandwidth of filtering. This of course does not decrease any losses just extends HF performance.
This is often but not always followed by regulators. These started out as simple gas discharge regulators, changing to discrete transistor designs, followed by integrated versions. Then WJ and others looked at if better could be built and the reign of super regulators was born. Then some idiot did a tutorial on shunt regulators and they began to reappear. (Gas discharge regulators are shunts.)
Now I should point out a shunt regulator is not the same as an active capacitor emulator.
Since I am sure I could design an active device to be added to power supplies to reduce noise by 40 to 45 db (Oh I do better than that on my best power supplies) I see no reason to find the Bybee claims unreasonable here.
I certainly can measure all of the issues raised here. If folks tell me they can't hear them I also find that claim reasonable.
Let us be clear about the difference between a power amplifier power supply requirement and other lower level gear.
A high end power amplifier typically will have very large power supply filter capacitor banks. These will reduce ripple and in theory allow large peak currents. In practice if coupled with an efficient transformer they will place very large peak current demands on the AC power line. That is because low ripple also means less available time at the AC cycle peak to charge. This higher demand current makes issues such as AC line resistance much more important. A typical very good AC line and power cord will have .2 ohms of resistance. If you consider that 100mv of ripple on a 50 volt supply is 500 to one, you begin to understand why .2 ohms can be a big deal. (The inductance in the transformer reduces this 500 x peak)
The front end of a power amplifier has very similar needs for stable voltages as do the other small signal pieces of gear.
So the first issue for small signal gear is what happens to the internal power supply when the power amplifier is clipping the AC power line voltage peaks?
Almost any regulator will solve the voltage issue, but narrow current draw pulses also make noise!
Some will place a capacitor across the incoming AC line to reduce some of the line noise issues.
It is obvious that some of the noise makes it past the power supply transformer. Shields will of course reduce noise but never eliminate it. As most who have encountered an internal transformer shield know there is an issue of how you ground it and does the safety ground reduce or increase the noise.
After the transformer are of course rectifier diodes. The wise high end power supply designer may use special diodes to reduce switching noise, some may even place small resistors here to smooth and stretch currents, some even go so far as to use parallel diode resistor combinations.
The filter capacitors for those who have looked have internal resistance and inductance. This limits their effective filtering action. Most will place smaller capacitors across larger ones to get greater bandwidth of filtering. This of course does not decrease any losses just extends HF performance.
This is often but not always followed by regulators. These started out as simple gas discharge regulators, changing to discrete transistor designs, followed by integrated versions. Then WJ and others looked at if better could be built and the reign of super regulators was born. Then some idiot did a tutorial on shunt regulators and they began to reappear. (Gas discharge regulators are shunts.)
Now I should point out a shunt regulator is not the same as an active capacitor emulator.
Since I am sure I could design an active device to be added to power supplies to reduce noise by 40 to 45 db (Oh I do better than that on my best power supplies) I see no reason to find the Bybee claims unreasonable here.
I certainly can measure all of the issues raised here. If folks tell me they can't hear them I also find that claim reasonable.
Power Supplies Ain't Just Power Supplies ???.....
JC goes to great trouble to provide individual low noise regulation for every stage in some of his gear...there was discussion and measurements on how tuning LM317/337 3 pin regulators affects sound.
One would think that RC decoupled supplies would be fine.....perhaps not entirely so, and it appears that active PS regulators can be very audible even in opamp stages with high PSRR.
In opamps specified as high PSRR it is curious that they are so supply sonically dependent according to subjective reports.
Eric.
JC goes to great trouble to provide individual low noise regulation for every stage in some of his gear...there was discussion and measurements on how tuning LM317/337 3 pin regulators affects sound.
One would think that RC decoupled supplies would be fine.....perhaps not entirely so, and it appears that active PS regulators can be very audible even in opamp stages with high PSRR.
In opamps specified as high PSRR it is curious that they are so supply sonically dependent according to subjective reports.
Eric.
The question really is not whether you can hear "noise" as such through your speakers. The question is can you hear the effect of cleaning up power supply noise on the sound. I can, but you have already said you won't believe me. I will believe you if you say you can't hear it, but it seems to me you have made up your mind already that since you can't hear noise in the audio signal then you won't be able to detect any change in the sound.
Will you please stop your continued snooty references to "magic devices" I have already alluded to the type of device this is; this is an active noise reduction circuit, there is no magic claimed, you appear to be unable to get it, I can't do anything about that I am afraid.
Last edited:
I still think this music rail is somewhat interesting. (though I have no plans to buy one) I also think the marketing of it could be much more straightforward, without the FUD, but that seems to simply be the way bybee markets. So much secrecy...
I must also say that while other "products" bearing his name in the past have seemed so much snake oil.. that it may be hard for this to overcome that bias. However, if we step back a moment, and ask what if some other company offered this it may not bring as much derision? (though any other company might likely state more clearly what and how it works)
I don't think it is a good fit for pure DIY. We want to what and how it is working and to implement and tweak ourselves. And there are many ways to get superb power supply regulation and low noise, as well as good PSRR of the circuits themselves...
it's probably not worth the hype. And I'm skeptical at best that it would make any difference in a well designed quality piece of equipment. But I've also got little doubt that there are many products out in the world with sub-par power supply design that could certainly benefit from such an upgrade. There are other way to do so, this is just one more. (like Burson products? - never used them either, just sayin...)
as for audibility of noise? I basically agree with Robert F. It may not be possible to "hear" the noise. It is very likely that the benefits of removing the noise are audible. And of having a good low impedance power supply. Noise is certainly measurable. Effect on sound could be argued forever...
I must also say that while other "products" bearing his name in the past have seemed so much snake oil.. that it may be hard for this to overcome that bias. However, if we step back a moment, and ask what if some other company offered this it may not bring as much derision? (though any other company might likely state more clearly what and how it works)
I don't think it is a good fit for pure DIY. We want to what and how it is working and to implement and tweak ourselves. And there are many ways to get superb power supply regulation and low noise, as well as good PSRR of the circuits themselves...
it's probably not worth the hype. And I'm skeptical at best that it would make any difference in a well designed quality piece of equipment. But I've also got little doubt that there are many products out in the world with sub-par power supply design that could certainly benefit from such an upgrade. There are other way to do so, this is just one more. (like Burson products? - never used them either, just sayin...)
as for audibility of noise? I basically agree with Robert F. It may not be possible to "hear" the noise. It is very likely that the benefits of removing the noise are audible. And of having a good low impedance power supply. Noise is certainly measurable. Effect on sound could be argued forever...
It reduces the Z that the circuit (or as you say, the drain) "sees," like a buffer (or any good regulator stage), though as you say it cannot make the total power supply step response or "stiffness" and better than it is. the supply would still droop given a large load. However the circuit being powered would see a low impedance. (within limits - not sure if it is proper to call this local impedance?)
IIRC, the 'rail has a GND connection, so it's not a simple series element.
Since they're generally available, I can't wait to see a technical review of them with some real measurements.
according to the link posted earlier by tvi: "Bybee, John William"
It seems many are approaching this suspiciously (to say the least) based on Mr. Bybee's past history. (I am as well
) I would be much less suspicious if the marketing of it were more open... And if someone would just send some pictures of the innards. (though I'd not be surprised if they're potted) And maybe some spectrum/noise analyzer plots...I suppose as always the fear is for the product, if it as good as advertised, (Mr. Curl says it is) that it would then be ripped off and copied.
But it is patented apparently. (if that is the correct patent) and so the only real risk is from those who don't care about patents, and nothing will stop someone with the resources from reverse engineering it if they really think it is worthwhile to do so.
-CK
sure - thats the claim. But how?
John, what were the performance figures you got and what specific performance attributes tickled your fancy?
Consider a voltage source with a finite output impedance. If you want to reduce the impedance seen by what comes next there are two options: add a shunt impedance to ground, or add a series negative impedance. A combination of the two is possible. If this circuit reduces Z then that is how it does it.
If the patent is 7521990, then the embodiment for DC rail filtering appears to be this:
So basically take the rail voltage, run it through a filter, and buffer that reference voltage to a follower. Haven't guys been doing this for 40+ years already? (BTW, X2 is resistor divider or resistor-zener, and X3 is RC network.)
Maybe the AC embodiments of the patent are novel. I'm not clear what's new here.
So basically take the rail voltage, run it through a filter, and buffer that reference voltage to a follower. Haven't guys been doing this for 40+ years already? (BTW, X2 is resistor divider or resistor-zener, and X3 is RC network.)
Maybe the AC embodiments of the patent are novel. I'm not clear what's new here.
Attachments
Last edited:
Hi DF96,
However, it's still up to the discretion of the examiner ...
If it's a variation on a circuit by Wenzel, it probably works very well. My own work is in the direction to remove the HF components and glitches. The LF stuff isn't going to be a problem as the junction capacitances of semiconductors aren't large enough to be a factor around 120 Hz or so.
-Chris
Yup, that's correct.
However, it's still up to the discretion of the examiner ...
If it's a variation on a circuit by Wenzel, it probably works very well. My own work is in the direction to remove the HF components and glitches. The LF stuff isn't going to be a problem as the junction capacitances of semiconductors aren't large enough to be a factor around 120 Hz or so.
-Chris
Hi All,
After looking around for the old article, I found it hiding here with it's creator(s). See the article: Reducing Regulator Noise, or "Finesse Voltage Regulator Noise!" (as I have it here). The host is Wenzel Associates, Inc. Very kind to post some of their work for everyone.
This circuit is real, it works. You'll notice that the first implementation with a single transistor claims a noise reduction in the same ballpark as Bybee claims.
The big question now is, did Mr. Bybee rip off Wenzel, or have they come up with the same idea on their own and failed to research properly. This may sink any patent claims - depending on what it is exactly. Another point could be made that now the invention has patent protection, and plenty of cash to enforce it, the schematic could be published without fear of infringement. Fears of the Far East copying it are null as soon as the first one is sold. After all, it's less trouble to copy a working, produced unit than it is to create from a schematic.
-Chris
After looking around for the old article, I found it hiding here with it's creator(s). See the article: Reducing Regulator Noise, or "Finesse Voltage Regulator Noise!" (as I have it here). The host is Wenzel Associates, Inc. Very kind to post some of their work for everyone.
This circuit is real, it works. You'll notice that the first implementation with a single transistor claims a noise reduction in the same ballpark as Bybee claims.
The big question now is, did Mr. Bybee rip off Wenzel, or have they come up with the same idea on their own and failed to research properly. This may sink any patent claims - depending on what it is exactly. Another point could be made that now the invention has patent protection, and plenty of cash to enforce it, the schematic could be published without fear of infringement. Fears of the Far East copying it are null as soon as the first one is sold. After all, it's less trouble to copy a working, produced unit than it is to create from a schematic.
-Chris
Last edited:
That a National part there Simon?
My error in the host in my previous post. I found the article through the Wenzel site, I have no idea if they are related or not. The host is actually "http://www.techlib.com".
Still, really nice of them, and the concept works specifically on HF noise. I guess you could bend this to any frequency range you want using filters. I was going to remove any 10 MHz off the supply in a distribution amp, but then I figured it might be better to leave it broad band, but higher frequency AC only. It's that high frequency stuff that tends to pass through transistors and regulator ICs.
-Chris
Edit
Hmmm, this patent, Trimble Navigation US 5408193 (A), seems to be pretty close. Year is 1995.
My error in the host in my previous post. I found the article through the Wenzel site, I have no idea if they are related or not. The host is actually "http://www.techlib.com".
Still, really nice of them, and the concept works specifically on HF noise. I guess you could bend this to any frequency range you want using filters. I was going to remove any 10 MHz off the supply in a distribution amp, but then I figured it might be better to leave it broad band, but higher frequency AC only. It's that high frequency stuff that tends to pass through transistors and regulator ICs.
-Chris
Edit
Hmmm, this patent, Trimble Navigation US 5408193 (A), seems to be pretty close. Year is 1995.
Last edited:
DF96, getting a patent granted and having it survive litigation or re-examination are two different things. Unlike the European system, the examiner is the only one (other than the applicant) who can cite prior art. And most examiners will only look at patent prior art, not places such as old books or app notes. This creates much opportunity for patent attorneys to generate litigation revenue (we do not have a loser-pays system, so the lawyers always win).
That confirms something I had imagined, that patent examiners only search patents for prior art. I guess they assume that patents are so wonderful and important that any worthwhile prior art would have been patented, rather than merely published in textbooks or magazine articles. I think someone (my PhD supervisor?) once told me that it is fairly safe to ignore US patents, as few of them would be patentable in Europe.
I hate to intrude here, but the design is real, reasonably sophisticated, and it works. I just tested a pair this week, and it reduced the self noise of the supplies to below what the ST analyzer can measure. I will have to change test equipment to measure properly.
Chipping at the design concepts, especially after looking at the patent disclosure, is typical of people who don't like to give credit to others.
Chipping at the design concepts, especially after looking at the patent disclosure, is typical of people who don't like to give credit to others.
- Status
- Not open for further replies.