Yup, sounds like you've got the idea.
Assuming that a bridge were laid out this way:
+~
~-
then caps would go from + to ~, from ~ to -, from - to ~, and from ~ to +, chasing each other clockwise around the outside of the square, with none taking shortcuts across the middle.
Grey
Assuming that a bridge were laid out this way:
+~
~-
then caps would go from + to ~, from ~ to -, from - to ~, and from ~ to +, chasing each other clockwise around the outside of the square, with none taking shortcuts across the middle.
Grey
High Speed Diodes
High speed diodes are a very audible improvement on most of the equipment I have modified and built. I like the IR Hexfreds and
the IXYS diodes from Digikey. The GI4396 are very cost effective for small supplies with 1000uF or less filter caps. I also like the GI851s for filter caps under 5000uF. Hanging caps across diodes
is not a slam dunk. The lead inductance will resonant with the capacitor at R.F. frequencies and can make things worse. I have seen diodes damped with RC circuits, but the cap and resistor type and values are critical and beyond the scope of most constructors. A good link on the subject can be found at the General Semiconductor web site at: www.gensemi.com/appnotespdf/quik108.pdf
H.H.
High speed diodes are a very audible improvement on most of the equipment I have modified and built. I like the IR Hexfreds and
the IXYS diodes from Digikey. The GI4396 are very cost effective for small supplies with 1000uF or less filter caps. I also like the GI851s for filter caps under 5000uF. Hanging caps across diodes
is not a slam dunk. The lead inductance will resonant with the capacitor at R.F. frequencies and can make things worse. I have seen diodes damped with RC circuits, but the cap and resistor type and values are critical and beyond the scope of most constructors. A good link on the subject can be found at the General Semiconductor web site at: www.gensemi.com/appnotespdf/quik108.pdf
H.H.
costs and quality of power supplies
I knew nothing about electronics but I decided to learn and eventually made a SOZ. It seems that most agree that a properly sized choke and capacitor filter sounds better. The major draw back is its larger size and cost. But, I concluded that it really does not cost that much more. As a a beginner any regulated, capacitor multiplier or any other active device based power supply out there would require quite a bit of time to learn about . Then spend a lot of time adapting it to my needs, tweaking it and testing it. And finally take longer to construct. A simple unregulated choke supply of any type is fairly easy to learn and build. That means I save a lot of time. As they say "time is money". Plus the most important feature is many agree it sounds the best. One reason to go DIY is that you can build the best with better parts for much less than the same product would cost if offered by a manufacturer. Good air core inductors with up to 10 gauge wire can be bought direct from www.Solen.ca They are in Canada and are the same people who make Solen caps.
I knew nothing about electronics but I decided to learn and eventually made a SOZ. It seems that most agree that a properly sized choke and capacitor filter sounds better. The major draw back is its larger size and cost. But, I concluded that it really does not cost that much more. As a a beginner any regulated, capacitor multiplier or any other active device based power supply out there would require quite a bit of time to learn about . Then spend a lot of time adapting it to my needs, tweaking it and testing it. And finally take longer to construct. A simple unregulated choke supply of any type is fairly easy to learn and build. That means I save a lot of time. As they say "time is money". Plus the most important feature is many agree it sounds the best. One reason to go DIY is that you can build the best with better parts for much less than the same product would cost if offered by a manufacturer. Good air core inductors with up to 10 gauge wire can be bought direct from www.Solen.ca They are in Canada and are the same people who make Solen caps.
Regulated Supples
Heres a good supply topology in the link. I have built simular and it sounds great. I like LC supplies also and even better is to follow the LC filter with the souce follower circuit. You can reference the gate to the unregulated input voltage which will give you the gate to source voltage across the drain to source. Power Supply design is one of the most important influences in the sound of an audio circuit and well worth the effort.
Single Ended Pure Class A Power Follower Amplifier
http://www.geocities.com/ResearchTriangle/8231/my/
H.H.
Heres a good supply topology in the link. I have built simular and it sounds great. I like LC supplies also and even better is to follow the LC filter with the souce follower circuit. You can reference the gate to the unregulated input voltage which will give you the gate to source voltage across the drain to source. Power Supply design is one of the most important influences in the sound of an audio circuit and well worth the effort.
Single Ended Pure Class A Power Follower Amplifier
http://www.geocities.com/ResearchTriangle/8231/my/
H.H.
Going back to Variac;
I much prefer RC snubbers over just capacitors across
each diode leg. In measurements, I have seen the
RF emission simply shifted down in frequency, often at
a larger amplitude through the resonant LC network
of transformer and wiring.
A snubber would typically consist of a resistor and cap in
series, usually 5-10 ohms and .01 to .1 uF.
I much prefer RC snubbers over just capacitors across
each diode leg. In measurements, I have seen the
RF emission simply shifted down in frequency, often at
a larger amplitude through the resonant LC network
of transformer and wiring.
A snubber would typically consist of a resistor and cap in
series, usually 5-10 ohms and .01 to .1 uF.
Referring back to Variac and NP
There was an excellent article in Audio Electronics on how to calculate optimal values of:
1. Capacitor (to lower resonant frequency)
2. Simultaneous snubber (to remove as much as possible of this new lowered resonant mode)
Only problem is that you need to measure the resonant frequency and set it up for that unit (or series of units) -- which means an oscilloscope.
Very very nice. Good theory, and easy steps to move ahead.
Probably extremely effective too.
Why don't manufacturers use this method (shameless plug on the Pass Labs forum 🙂 )
Petter
There was an excellent article in Audio Electronics on how to calculate optimal values of:
1. Capacitor (to lower resonant frequency)
2. Simultaneous snubber (to remove as much as possible of this new lowered resonant mode)
Only problem is that you need to measure the resonant frequency and set it up for that unit (or series of units) -- which means an oscilloscope.
Very very nice. Good theory, and easy steps to move ahead.
Probably extremely effective too.
Why don't manufacturers use this method (shameless plug on the Pass Labs forum 🙂 )
Petter
Hi all
I also did quite some experiments with powers supplys and here is what I found.
I have a class AB amp, so an LC-filter doesn't work that great. As I work with high efficiency speakers, I do not use alot of power. So for me a capacitance multiplier with voltage refenence (zener diode with very high quality cap in parallell to filter noise from zener = voltage regulator) works best. For transistors I use the same Darlington combination(driver transistor and power transistor)as the amp itself. This gives me the advantage that I can buy a few more, so I can match the outputtransistors more closely (and if their is such a thing as a "sonic character of transistors", then it is at least twice the same character🙂 ).
As for the rectifier diodes, I also prefer the modern UltraFast, soft recovery ones. It is quite true we do not need the "ultra fast" in our hifi-application, but we defenately do need the "soft recovery". As these diodes are normally designed for switched power supplys (where they need both characteristics), we usually can't find "soft recovery" without "ultra fast", so there is sometimes the misconception we need an "ultra fast" diode....
Caps across diodes is IMO indeed better then just plain diodes, but worse then soft recovery types. The caps do indeed filter the switching spikes, but they also form a direct path for HF-noise coming from the power supply lines (all transfos have an interwinding capacitance, so with this interwinding capacitance and the cap across the diodes you have a direct path for HF-noise to your power amp). This of course can be solved with a small choke (but even small chokes for high current are expensive), so choke + caps + normal diodes are more expensive then an "soft recovery" rectifier brigde.
More complexe regulators (like Jung regulator) have even a better regulation, but can't deliver as high current and are prone to oscillation. The also use alot of feedback and have a smaller bandwith, so these are in my opinion not as ideal.
Another thing to consider is two built two identical positive power supply and afterwards reference them together as a positive and negative voltage (by connecting one of the positive voltages to the ground of the other supply). This way you have a "balanced" power supply, in with your ground plain only sees "signal curren".
I will try and make a circuitdiagram of the regulator I now uses, and try and post it on this forum. But as I am quitebusy the next few days, it may take a while.
Greetz
Wim
I also did quite some experiments with powers supplys and here is what I found.
I have a class AB amp, so an LC-filter doesn't work that great. As I work with high efficiency speakers, I do not use alot of power. So for me a capacitance multiplier with voltage refenence (zener diode with very high quality cap in parallell to filter noise from zener = voltage regulator) works best. For transistors I use the same Darlington combination(driver transistor and power transistor)as the amp itself. This gives me the advantage that I can buy a few more, so I can match the outputtransistors more closely (and if their is such a thing as a "sonic character of transistors", then it is at least twice the same character🙂 ).
As for the rectifier diodes, I also prefer the modern UltraFast, soft recovery ones. It is quite true we do not need the "ultra fast" in our hifi-application, but we defenately do need the "soft recovery". As these diodes are normally designed for switched power supplys (where they need both characteristics), we usually can't find "soft recovery" without "ultra fast", so there is sometimes the misconception we need an "ultra fast" diode....
Caps across diodes is IMO indeed better then just plain diodes, but worse then soft recovery types. The caps do indeed filter the switching spikes, but they also form a direct path for HF-noise coming from the power supply lines (all transfos have an interwinding capacitance, so with this interwinding capacitance and the cap across the diodes you have a direct path for HF-noise to your power amp). This of course can be solved with a small choke (but even small chokes for high current are expensive), so choke + caps + normal diodes are more expensive then an "soft recovery" rectifier brigde.
More complexe regulators (like Jung regulator) have even a better regulation, but can't deliver as high current and are prone to oscillation. The also use alot of feedback and have a smaller bandwith, so these are in my opinion not as ideal.
Another thing to consider is two built two identical positive power supply and afterwards reference them together as a positive and negative voltage (by connecting one of the positive voltages to the ground of the other supply). This way you have a "balanced" power supply, in with your ground plain only sees "signal curren".
I will try and make a circuitdiagram of the regulator I now uses, and try and post it on this forum. But as I am quitebusy the next few days, it may take a while.
Greetz
Wim
Snubbers
Thanks for the heads up. Which issue of A.E. Many audio designers don't know about this problem. A lot of them are going to high speed soft recovery diodes to address this though without fully understanding what they are fixing. You have to remember parts cost money and add at least a X6 addition to the retail price ( estimate for one of my audio designer friends). Most of my knowledge on this comes from switching power supply design and EMI reduction work. Don't for get to use smallest caps nd shortest leads possible for these circuits. A 600V polystyrene cap is not what you want for this one. A 100 volt stacked metalized foil polyester would probably better for this application.
Thanks,
H.H.
Thanks for the heads up. Which issue of A.E. Many audio designers don't know about this problem. A lot of them are going to high speed soft recovery diodes to address this though without fully understanding what they are fixing. You have to remember parts cost money and add at least a X6 addition to the retail price ( estimate for one of my audio designer friends). Most of my knowledge on this comes from switching power supply design and EMI reduction work. Don't for get to use smallest caps nd shortest leads possible for these circuits. A 600V polystyrene cap is not what you want for this one. A 100 volt stacked metalized foil polyester would probably better for this application.
Thanks,
H.H.
- Status
- Not open for further replies.
- Home
- Amplifiers
- Pass Labs
- Capacitor multiplier supply for ALEPH 3