peter
you did a test before which used your orginal single transformer verses two dual transforms - did they use the same diodes?
i say this because this may have been done before you realised the difference between the different makes of diodes.
you did a test before which used your orginal single transformer verses two dual transforms - did they use the same diodes?
i say this because this may have been done before you realised the difference between the different makes of diodes.
The straight dope on doping diodes
I am not even going to claim to be semiconductor expert. My old boss used to come in my cubicle and try explain this stuff to me to help me design fast lightning protection circuits. He had some background in semiconductors. He worked at Texas Instruments way back. He claims to have been one of the witnesses to sign Jack Kirby's Lab notebook describing the idea for the first Integrated Circuit. Lab Notebooks are often used as documents of proof in patent applications and require witnesses signatures.
I think you guys lay awake at night to think up ways to get me to embarrass myself. I do have a few links that cover basic semiconductor theory and one that actually discusses the doping and fabrication methods for fast recovery epitaxial diodes.
Introductory level
http://hyperphysics.phy-astr.gsu.edu/hbase/solids/intrin.html#c1
http://hyperphysics.phy-astr.gsu.edu/hbase/solids/sili2.html#c1
http://hyperphysics.phy-astr.gsu.edu/hbase/solids/intrin.html#c2
http://hyperphysics.phy-astr.gsu.edu/hbase/solids/dope.html#c2
http://hyperphysics.phy-astr.gsu.edu/hbase/solids/pnjun.html#c1
http://hyperphysics.phy-astr.gsu.edu/hbase/solids/diod.html#c3
A very good overview on semiconductor theory with more detail:
http://www.hep.fsu.edu/~wahl/Quarknet/summer2001/conductivity.pdf
Doping of Fast Recovery Epitaxial Diodes
http://www.ixys.com/t052599a.pdf
Slow to Recover, Taxable, and hoping to Die old,
Fred
http://www.straightdope.com/
Dieckmann/ Is starting to feel like the Cecil Adams of High End Audio
I am not even going to claim to be semiconductor expert. My old boss used to come in my cubicle and try explain this stuff to me to help me design fast lightning protection circuits. He had some background in semiconductors. He worked at Texas Instruments way back. He claims to have been one of the witnesses to sign Jack Kirby's Lab notebook describing the idea for the first Integrated Circuit. Lab Notebooks are often used as documents of proof in patent applications and require witnesses signatures.
I think you guys lay awake at night to think up ways to get me to embarrass myself. I do have a few links that cover basic semiconductor theory and one that actually discusses the doping and fabrication methods for fast recovery epitaxial diodes.
Introductory level
http://hyperphysics.phy-astr.gsu.edu/hbase/solids/intrin.html#c1
http://hyperphysics.phy-astr.gsu.edu/hbase/solids/sili2.html#c1
http://hyperphysics.phy-astr.gsu.edu/hbase/solids/intrin.html#c2
http://hyperphysics.phy-astr.gsu.edu/hbase/solids/dope.html#c2
http://hyperphysics.phy-astr.gsu.edu/hbase/solids/pnjun.html#c1
http://hyperphysics.phy-astr.gsu.edu/hbase/solids/diod.html#c3
A very good overview on semiconductor theory with more detail:
http://www.hep.fsu.edu/~wahl/Quarknet/summer2001/conductivity.pdf
Doping of Fast Recovery Epitaxial Diodes
http://www.ixys.com/t052599a.pdf
Slow to Recover, Taxable, and hoping to Die old,
Fred
http://www.straightdope.com/
Dieckmann/ Is starting to feel like the Cecil Adams of High End Audio
fezz said:
I even compared them again today. Both supplies use the same diodes. One employs a single 400VA Plitron toroid, the other 2 x 200VA EI surplus transformers. Plitron seems to sound better, kind of cleaner.
Re: Whats The Diff ?.
International Rectifier has very good tutorials on the operation of the various devices that they sell. You can find their tutorial on high speed / soft recovery diodes here:
http://www.irf.com/technical-info/whitepaper/murdiodes.pdf
The key point is that the way one makes a diode "soft" is to reduce the lifetime of minority carriers. Typically this is done by diffusing small amounts of metal into the silicon or by damaging the silicon slightly with electron radiation. IR claims that they get better results through the use of platinum as the lifetime killer instead of the more common gold.
---Gary
mrfeedback said:
International Rectifier has very good tutorials on the operation of the various devices that they sell. You can find their tutorial on high speed / soft recovery diodes here:
http://www.irf.com/technical-info/whitepaper/murdiodes.pdf
The key point is that the way one makes a diode "soft" is to reduce the lifetime of minority carriers. Typically this is done by diffusing small amounts of metal into the silicon or by damaging the silicon slightly with electron radiation. IR claims that they get better results through the use of platinum as the lifetime killer instead of the more common gold.
---Gary
Re:Re:What's the diff?
Hi,
Indeed, a clever idea...by make the semi-conductor slightly more conductive the diode conducts a bit more, is slightly less effecient at rectifying but it does the job nicely.
Cheers,😉
Hi,
Indeed, a clever idea...by make the semi-conductor slightly more conductive the diode conducts a bit more, is slightly less effecient at rectifying but it does the job nicely.
Cheers,😉
Frank,
Actually it doesn't work quite the way you think.
The amount of metal added is quite small, on the order of 1e10atoms/cm-3 to 1e12atoms/cm-3, which works out to be between 2e-9 % and 2e-11 %. At those levels, it has no effect on the conductivity of the material. It does increase reverse bias junction leakage currents and decreases the recombination lifetime of minority carriers by introducing deep level traps into the bandgap of the silicon. Its these effects that give the soft recovery characteristic, not a change in conductivity.
---Gary
Actually it doesn't work quite the way you think.
The amount of metal added is quite small, on the order of 1e10atoms/cm-3 to 1e12atoms/cm-3, which works out to be between 2e-9 % and 2e-11 %. At those levels, it has no effect on the conductivity of the material. It does increase reverse bias junction leakage currents and decreases the recombination lifetime of minority carriers by introducing deep level traps into the bandgap of the silicon. Its these effects that give the soft recovery characteristic, not a change in conductivity.
---Gary
DIODES.
Hi,
Deep sigh...I agree with you entirely, guess I over "Reader Digest'ed it?
I know what you mean though, the same is happening in the computor industry, my livelyhood you know...😎
Cheers,😉
Hi,
Deep sigh...I agree with you entirely, guess I over "Reader Digest'ed it?
I know what you mean though, the same is happening in the computor industry, my livelyhood you know...😎
Cheers,😉
THE SOUND OF DIODES.
Hi,
Has any of you tested the relationship between voltage level and diode sound?
Just wondering...
Cheers,😉
Hi,
Has any of you tested the relationship between voltage level and diode sound?
Just wondering...
Cheers,😉
Re: DIODES.
Frank,
Sorry if I got overly pedantic. The desire to explain things in terms that a non-professional can understand is admirable. The only problem is that there is a fine line between simplification and distortion of the basic mechanisms. I know you weren't trying to do this, but we have plenty of energetic folks around here who will actually take what we say at face value and try to measure the effects we describe. Since I do this stuff for a living, I thought I'd pass along a few of the details for those interested.
---Gary
fdegrove said:
Frank,
Sorry if I got overly pedantic. The desire to explain things in terms that a non-professional can understand is admirable. The only problem is that there is a fine line between simplification and distortion of the basic mechanisms. I know you weren't trying to do this, but we have plenty of energetic folks around here who will actually take what we say at face value and try to measure the effects we describe. Since I do this stuff for a living, I thought I'd pass along a few of the details for those interested.
---Gary
Hoo Ray I am off the hook!
"Sorry if I got overly pedantic."
Pendance away Gary! I knew we wound flush a semiconducter guru out eventually.
The http://www.irf.com/technical-info/whitepaper/murdiodes.pdf
link was the first one I put in this thread. Any more insight would be welcome. The question about voltage rating raised is a good one. Most of the FREDs are rated for 600 volts to be used on the AC line input sourced switching power supplies I believe. I also recall seeing a few FREDs that were several series connected diodes which I believe had something to do with the recovery profile vs. reverse breakdown voltage. the IXYS appnote I included seemed to indicate they belived that gold doping had softer recovery characteristics than platinum doping.
There are times I almost think
I am not sure of what I absolutely know.
Very often find confusion
In conclusion I concluded long ago
In my head are many facts
That, as a student, I have studied to procure,
In my head are many facts..
Of which I wish I was more certain I was sure!
But...Is a puzzlement!
"Sorry if I got overly pedantic."
Pendance away Gary! I knew we wound flush a semiconducter guru out eventually.
The http://www.irf.com/technical-info/whitepaper/murdiodes.pdf
link was the first one I put in this thread. Any more insight would be welcome. The question about voltage rating raised is a good one. Most of the FREDs are rated for 600 volts to be used on the AC line input sourced switching power supplies I believe. I also recall seeing a few FREDs that were several series connected diodes which I believe had something to do with the recovery profile vs. reverse breakdown voltage. the IXYS appnote I included seemed to indicate they belived that gold doping had softer recovery characteristics than platinum doping.
There are times I almost think
I am not sure of what I absolutely know.
Very often find confusion
In conclusion I concluded long ago
In my head are many facts
That, as a student, I have studied to procure,
In my head are many facts..
Of which I wish I was more certain I was sure!
But...Is a puzzlement!
Attachments
Re: Hoo Ray I am off the hook!
Fred,
I guess we should call this a "flush wound"?
---Gary
p.s. my apologies if people don't like bad puns
Fred,
I guess we should call this a "flush wound"?
---Gary
p.s. my apologies if people don't like bad puns
Diode physics
Fred,
I have to confess that I didn't read the entire thread, thus I missed your initial posting of the IR reference. Regarding insights into the design of the diodes, I can try but its not my specialty. Designing high voltage devices is a rather arcane specialty with a lot of empirical know-how that isn't very well documented. The devil is in the details and I can only give educated guesses as to what these folks are really doing.
The trick in designing high voltage devices is to control the way the voltage spreads itself across the silicon. The diode doesn't care about the voltage, only the electric field which is voltage divided by the physical distance over which it is applied. Thus a high voltage diode is designed to insure that the voltage is applied over a large enough piece of material that the breakdown field of the silicon and/or silicon dioxide insulator is not exceeded. The tradeoff is doing this without adding too much resistance thus limiting the current capability of the diode.
Anyway, thats a start. Does that help?
---Gary
Fred,
I have to confess that I didn't read the entire thread, thus I missed your initial posting of the IR reference. Regarding insights into the design of the diodes, I can try but its not my specialty. Designing high voltage devices is a rather arcane specialty with a lot of empirical know-how that isn't very well documented. The devil is in the details and I can only give educated guesses as to what these folks are really doing.
The trick in designing high voltage devices is to control the way the voltage spreads itself across the silicon. The diode doesn't care about the voltage, only the electric field which is voltage divided by the physical distance over which it is applied. Thus a high voltage diode is designed to insure that the voltage is applied over a large enough piece of material that the breakdown field of the silicon and/or silicon dioxide insulator is not exceeded. The tradeoff is doing this without adding too much resistance thus limiting the current capability of the diode.
Anyway, thats a start. Does that help?
---Gary
high voltage
The IXYS app note also describes the use of guard rings in the metalization layer to reduce the electric field strength to prevent voltage break-down.
I believe there is a realtionship between the reverse breakdown voltage and the recovery characteristics. The IXYS DSEP 9-06CR is actually three 200 volt diodes in series to acheive low tRR with soft recovery characteristics. The three foward drops will most likely make this unsuitable for power amplifier supplies due to the voltage loss.
http://www.ixys.com/l356.pdf
http://www.ixys.com/t03232ka.pdf
The IXYS app note also describes the use of guard rings in the metalization layer to reduce the electric field strength to prevent voltage break-down.
I believe there is a realtionship between the reverse breakdown voltage and the recovery characteristics. The IXYS DSEP 9-06CR is actually three 200 volt diodes in series to acheive low tRR with soft recovery characteristics. The three foward drops will most likely make this unsuitable for power amplifier supplies due to the voltage loss.
http://www.ixys.com/l356.pdf
http://www.ixys.com/t03232ka.pdf
Re: high voltage
Fred,
I can help tell you how these things are built, but like everyone else, I'm not sure I know which ones sound better. Or more precisely, why the ones that people think sound better actually sound better. The earlier posts about snubbers and critically damping resonances that could be excited by diode voltage spikes are right on the money. I tend to favor schottky diodes since they have no reverse recovery spike. Then I like to put a small RC network between the transformer and the diodes to decrease the likelyhood of exciting a transformer resonance. Typcally I put a 0.01uf cap across the transformer leads, small resistors (<10ohm) in series with each lead and then another 0.01uf cap on the other side of the resistors where they go into the diode bridge. This seems to quiet things down nicely. Careful measurement would allow a more optimized network but this seems to work reasonably well.
Regarding the IXYS DSEP 9-06CR, you are right that these have a pretty high forward voltage drop, consistent with them being 3 diodes in series. Still, 3v is not a huge loss and one can overspec the power transformer to accomodate the extra 2v drop.
---Gary
Fred Dieckmann said:
Fred,
I can help tell you how these things are built, but like everyone else, I'm not sure I know which ones sound better. Or more precisely, why the ones that people think sound better actually sound better. The earlier posts about snubbers and critically damping resonances that could be excited by diode voltage spikes are right on the money. I tend to favor schottky diodes since they have no reverse recovery spike. Then I like to put a small RC network between the transformer and the diodes to decrease the likelyhood of exciting a transformer resonance. Typcally I put a 0.01uf cap across the transformer leads, small resistors (<10ohm) in series with each lead and then another 0.01uf cap on the other side of the resistors where they go into the diode bridge. This seems to quiet things down nicely. Careful measurement would allow a more optimized network but this seems to work reasonably well.
Regarding the IXYS DSEP 9-06CR, you are right that these have a pretty high forward voltage drop, consistent with them being 3 diodes in series. Still, 3v is not a huge loss and one can overspec the power transformer to accomodate the extra 2v drop.
---Gary
DIODES.
Hi,
Indeed that is my experience too, I can live with a little more voltage drop.
In tube designs the 1N4007 and its equivalents are rather popular and to make these a bit "stronger" (read to make them withstand higher inrush current) often two of them are put in series.
Regarding snubbers, I hardly ever need to use them on my (highish + 300V) voltage circuits, I always specify mains transformers with an electrostatic shield and find it very effective at keeping out mainsborne crap.
Shottky diodes are the most advanced and are my fav diodes too, not too many handle high voltages and high current simultaneously though.
Yup.That helps a lot too.
Plus I often add a varistor on the primary side too limit inrush current.
Cheers,😉
Hi,
Indeed that is my experience too, I can live with a little more voltage drop.
In tube designs the 1N4007 and its equivalents are rather popular and to make these a bit "stronger" (read to make them withstand higher inrush current) often two of them are put in series.
Regarding snubbers, I hardly ever need to use them on my (highish + 300V) voltage circuits, I always specify mains transformers with an electrostatic shield and find it very effective at keeping out mainsborne crap.
Shottky diodes are the most advanced and are my fav diodes too, not too many handle high voltages and high current simultaneously though.
Yup.That helps a lot too.
Plus I often add a varistor on the primary side too limit inrush current.
Cheers,😉
The serial....diodes....
Wrong...sorry!!
Two diode in series is for them to withstand higher peek voltages...the current handling capacity is the same!!!😉
Wrong...sorry!!
Two diode in series is for them to withstand higher peek voltages...the current handling capacity is the same!!!😉
NEED A BONE TO PICK.
Hi,
Yes, but that's not what I do it for...reread or try it yourself.
Cheers,😉
Hi,
Yes, but that's not what I do it for...reread or try it yourself.
Cheers,😉
not to try to snub anyone
The series resistors and excellent idea. They will limit the peak current trough the diodes, minimizing reverse recovery effects, which are are function of the forward current through the diode. Series resistor will also from a low pass filter in conjunction with the caps across the diode and transformer side. Putting a cap directly across transformer winding will cause a relativly high Q resonance with the leakage inductance of the secondary. I have heard of people putting a 0.1uF to 1 uF across the transformer secondary. The best value sonically is said to be found experimentally. I believe this has to do with the exact frequency of this resonance and how the power supply circuit reacts to this resonance. A capacitor on the transformer sideof the diode bridge concerns me since it acts to provide higher peak forward current through the diode bridge and greater reverse recovery effects as a result.
I have just one word........damping. The use of crtical damping for these resonant circuits, from added capacitors and the existing transfomer inductance, seems to be a good idea and one with relatively little experimentation so far for audio power supplies.
GaryB said:
The series resistors and excellent idea. They will limit the peak current trough the diodes, minimizing reverse recovery effects, which are are function of the forward current through the diode. Series resistor will also from a low pass filter in conjunction with the caps across the diode and transformer side. Putting a cap directly across transformer winding will cause a relativly high Q resonance with the leakage inductance of the secondary. I have heard of people putting a 0.1uF to 1 uF across the transformer secondary. The best value sonically is said to be found experimentally. I believe this has to do with the exact frequency of this resonance and how the power supply circuit reacts to this resonance. A capacitor on the transformer sideof the diode bridge concerns me since it acts to provide higher peak forward current through the diode bridge and greater reverse recovery effects as a result.
I have just one word........damping. The use of crtical damping for these resonant circuits, from added capacitors and the existing transfomer inductance, seems to be a good idea and one with relatively little experimentation so far for audio power supplies.
Re: DIODES.
The peak current will be limited from this which is a good idea.
I think the use of snubbers across each diode will minimize RF hash from the diode reverse recovery effects from getting into the supply and audio circuits. Substitution of the 1N4947 for the 1N4005 and other standard recovery 1 Amp diodes below 600 volts is a pretty good sonic improment also. I have used these quite often on low current high voltage power supplies for tube circuits.
http://www.gensemi.com/pdfs/1n4933.pdf
fdegrove said:
The peak current will be limited from this which is a good idea.
fdegrove said:
I think the use of snubbers across each diode will minimize RF hash from the diode reverse recovery effects from getting into the supply and audio circuits. Substitution of the 1N4947 for the 1N4005 and other standard recovery 1 Amp diodes below 600 volts is a pretty good sonic improment also. I have used these quite often on low current high voltage power supplies for tube circuits.
http://www.gensemi.com/pdfs/1n4933.pdf