Steven,
you are of course right, my post was just a general note about chokes.
If we would like to "smoothen" out the current spike arising with the filtering capacitor then a single choke or differential choke would be the choice.
And all other...
And the diode issue, this thread was about diodes but steered in the end to cover something else.
Introducing a choke won't necessary address the problems with diodes, just add another ringing component.
A choke with couple of mH or so have probably also windings tight to each other which add a noticeable parasitic capacitance and the choke will ring with it's own frequency, eg just a new oscillating component.
The voltage change over the diode is the same even if the current spike can be decreased dramatically by adding a series choke, still waiting for a Mr's "simulation"!
Correct snubber and filtering with the focus on the diode issue is the only way!
PS. I haven't done it myself and would actually see any research covering diodes in mains PSU's, have anybody made current measurings of the diodes trr (di/dt) curve for a 50/60Hz rectified voltage?
The days quiz!
Just a thought, a fast diode has a trr of 20 nS or so, how much have the voltage changed at zero crossover (where the diode commutate) at the same time for a 50/60 Hz mains voltage with an output of +-50 V?
Cheers
you are of course right, my post was just a general note about chokes.
If we would like to "smoothen" out the current spike arising with the filtering capacitor then a single choke or differential choke would be the choice.
And all other...
And the diode issue, this thread was about diodes but steered in the end to cover something else.
Introducing a choke won't necessary address the problems with diodes, just add another ringing component.
A choke with couple of mH or so have probably also windings tight to each other which add a noticeable parasitic capacitance and the choke will ring with it's own frequency, eg just a new oscillating component.
The voltage change over the diode is the same even if the current spike can be decreased dramatically by adding a series choke, still waiting for a Mr's "simulation"!
Correct snubber and filtering with the focus on the diode issue is the only way!
PS. I haven't done it myself and would actually see any research covering diodes in mains PSU's, have anybody made current measurings of the diodes trr (di/dt) curve for a 50/60Hz rectified voltage?
The days quiz!
Just a thought, a fast diode has a trr of 20 nS or so, how much have the voltage changed at zero crossover (where the diode commutate) at the same time for a 50/60 Hz mains voltage with an output of +-50 V?
Cheers
hjelm said:
Load dependent, what is dependent of what?
Speaking of diodes, a 100/120 Hz current ripple still is a verrrrry slow change for a verry fast diode!
Days quiz number two, how much is the current 100 nS before it's zero?
Conditions:
50 Hz
+-50 V
10000 uF
5 A load
Think about it,
cheers!
I would say load dependent as in if you have no cap after the rectifier and some load the voltage vould have changed ~0.1uv
If there is a cap and it hasn't discharged at all the voltage change would be close to 0, depends on how realistic models you have.
Therefore i claim that it is load (and capacitor) dependent.
But on the other hand i have been known to be wrong before
.
If there is a cap and it hasn't discharged at all the voltage change would be close to 0, depends on how realistic models you have.
Therefore i claim that it is load (and capacitor) dependent.
But on the other hand i have been known to be wrong before
.I have done simulations with ideal diodes to calculate the correct choke size. I implemented these chokes into actual hardware. I am presently running my Penultimate Zen amps with a choke-loaded diode bridge.
Simulations are good for showing pretty pictures. Building and testing the design gives one the authority to talk about the subject. I've done both: I simulated the design and then I built it.
So now I'm going to tell you my results, as you requested. However, these are not simulated results, they are real-world results.
The hardware behaves similar to the simulation - lower peak currents, better power factor, and less ripple current in the capacitor. I'm using a smaller transformer, it runs cooler, and my filter capacitors don't work as hard. The overall benefit is less power supply hum, because of lower ripple currents. I haven't done a side-by-side listening tests, but I prefer the choke-loaded supply simply because of less noise.
The story so far
From this thread I have excerpted what I feel are relevant passages relating to specific diodes and bridges that seem to be 'audiophile-approved.'
It sounds like OnSemi's MSR1560G 15A 600V ultrasoft recovery diode (~$1.30) is the best bang for the buck to make a high power bridge.
It also seems that there aren't any high power soft recovery bridges worth using, as Fairchild Stealths are out (edgy per PD) and the IXYS VBE 17-06NO7 27A 600V bridge (~$20), while being a FRED (fast recovery) isn't 'soft recovery.' What have I missed?
Peter Daniel:
I used regular bridges in my Aleph5, because at the time I wasn't aware of a difference thay make. MUR860 might be a bit small for that application, but you can either parallel them (but again I don't know how it changes sonics), otherwise you could use MUR1560 as I remember they sounded pretty good too. If you need more softness and presence in vocals, MUR1520 is a good candidate. Don't use diodes from Fairchild as they are very edgy and have glare. I found Motorola and On Semi parts better sounding.
In my AlephX I used IR 30ETH06 diodes, but again comparing 15ETH06 in my GC I found MUR better.
I just receive the MSR860 diodes today. I compared them to MUR860 and I must say that it seems like an improvement. More 3 dimentional sound, more detail in treble, more musical. MUR seems to be a bit flat sounding, while MSR gives more music. If I can complain about something, it would be a sort of artifact in female voice. I don't know what it really is, but it seems like some sort of haziness or glare, but nothing really big. I will see how it sounds after breaking in, but it looks very promising. Certainly the best alternative to MUR860 I checked so far.
Fred Dieckmann:
Substitution of the 1N4947 for the 1N4005 and other standard recovery 1 Amp diodes below 600 volts is a pretty good sonic improment also.
Elso Kwak:
Now as a experiment I replaced all Schottky's by MUR120 high speed soft recovery.
It is a clear improvement! These powersupplyrectifiers have a profound influence on the sound of female voices. Soprano's and flutes and the like. Sounds like some resonance effect to my ears.
So I tempted to state that the soft recovery issue is more important than the high speed.
In the Rectifier Applications Handbook from Onsemi (HB214/D) I found that Schottky's do exhibit ringing!
Onsemi has now a few ultrasoft recovery rectifiers: MSR1560, MSR860, MSRD620CT. The latter is 6A rating, the other even higher, but I will give it a try. Will also experiment with ferrite beads, RC-snubbers and small resistors in series with the diodes.
From this thread I have excerpted what I feel are relevant passages relating to specific diodes and bridges that seem to be 'audiophile-approved.'
It sounds like OnSemi's MSR1560G 15A 600V ultrasoft recovery diode (~$1.30) is the best bang for the buck to make a high power bridge.
It also seems that there aren't any high power soft recovery bridges worth using, as Fairchild Stealths are out (edgy per PD) and the IXYS VBE 17-06NO7 27A 600V bridge (~$20), while being a FRED (fast recovery) isn't 'soft recovery.' What have I missed?
Peter Daniel:
I used regular bridges in my Aleph5, because at the time I wasn't aware of a difference thay make. MUR860 might be a bit small for that application, but you can either parallel them (but again I don't know how it changes sonics), otherwise you could use MUR1560 as I remember they sounded pretty good too. If you need more softness and presence in vocals, MUR1520 is a good candidate. Don't use diodes from Fairchild as they are very edgy and have glare. I found Motorola and On Semi parts better sounding.
In my AlephX I used IR 30ETH06 diodes, but again comparing 15ETH06 in my GC I found MUR better.
I just receive the MSR860 diodes today. I compared them to MUR860 and I must say that it seems like an improvement. More 3 dimentional sound, more detail in treble, more musical. MUR seems to be a bit flat sounding, while MSR gives more music. If I can complain about something, it would be a sort of artifact in female voice. I don't know what it really is, but it seems like some sort of haziness or glare, but nothing really big. I will see how it sounds after breaking in, but it looks very promising. Certainly the best alternative to MUR860 I checked so far.
Fred Dieckmann:
Substitution of the 1N4947 for the 1N4005 and other standard recovery 1 Amp diodes below 600 volts is a pretty good sonic improment also.
Elso Kwak:
Now as a experiment I replaced all Schottky's by MUR120 high speed soft recovery.
It is a clear improvement! These powersupplyrectifiers have a profound influence on the sound of female voices. Soprano's and flutes and the like. Sounds like some resonance effect to my ears.
So I tempted to state that the soft recovery issue is more important than the high speed.
In the Rectifier Applications Handbook from Onsemi (HB214/D) I found that Schottky's do exhibit ringing!
Onsemi has now a few ultrasoft recovery rectifiers: MSR1560, MSR860, MSRD620CT. The latter is 6A rating, the other even higher, but I will give it a try. Will also experiment with ferrite beads, RC-snubbers and small resistors in series with the diodes.
More gooding sounding rectifiers
I'm resurrecting this thread and want to add a few more good sounding rectifiers. In my experience, only soft recovery rectifiers are good sounding. I found that it is also good to have the least amount of reverse leakage current.
Vishay HexFRED HFA25TB60PBF (600V 25A) and the 4A version are superb sounding, probably the best I've tried. They are ultrasoft recovery with very low reverse leakage current.
IXYS DSB40C common cathode soft recovery Schottky is also a good one, if you need low Vf low voltage for efficiency/less heat (5V application). But reverse leakage current is a bit high.
IXYS also makes some discrete soft recovery FRED rectifiers. One can look for them.
There is another brand that is very good sounding, but the manufacturer doesn't want to deal with individual customers, so I cannot give out their name. They have 35A 600V FWB that is soft recovery and low reverse leakage current. I was lucky to get it from someone close to the factory. I've not found any soft recovery high current FWB besides them. IXYS does not make soft recovery high current FWB, even places like PartsConnexion sells them (they sounded bright and shrilling IMO).
The datasheet has to explicitly state soft recovery, otherwise, it is not.
I'm resurrecting this thread and want to add a few more good sounding rectifiers. In my experience, only soft recovery rectifiers are good sounding. I found that it is also good to have the least amount of reverse leakage current.
Vishay HexFRED HFA25TB60PBF (600V 25A) and the 4A version are superb sounding, probably the best I've tried. They are ultrasoft recovery with very low reverse leakage current.
IXYS DSB40C common cathode soft recovery Schottky is also a good one, if you need low Vf low voltage for efficiency/less heat (5V application). But reverse leakage current is a bit high.
IXYS also makes some discrete soft recovery FRED rectifiers. One can look for them.
There is another brand that is very good sounding, but the manufacturer doesn't want to deal with individual customers, so I cannot give out their name. They have 35A 600V FWB that is soft recovery and low reverse leakage current. I was lucky to get it from someone close to the factory. I've not found any soft recovery high current FWB besides them. IXYS does not make soft recovery high current FWB, even places like PartsConnexion sells them (they sounded bright and shrilling IMO).
The datasheet has to explicitly state soft recovery, otherwise, it is not.
I would like to give my 2cents worth if I may.
To me the whole issue about fast diodes is not the correct approach. It is the soft recovery aspect of the diode that is the important characteristic. This is the characteristic that reduces EMI and kills the p-i-n diode effect of "normal" diodes. What also kills p-i-n diode effect is ca[acitance and this is where schottkey' come in as they have "high" capacitance over the junction. Zener diodes also do quite a good job in reducing p-i-n diode effects. These are the effects that can be measured and is quoted as EMI.
See article http://ixdev.ixys.com/DataSheet/4592a0fc-7250-493a-a3e4-c5c1c69dc02e.pdf
Vishay makes a fast soft recovery diode which I think is worth a try 20ETF... it is 20A rated and insulated.
Hope my response added some value.
Remember to sit back and listen to music and not only listen for faults ;-)
To me the whole issue about fast diodes is not the correct approach. It is the soft recovery aspect of the diode that is the important characteristic. This is the characteristic that reduces EMI and kills the p-i-n diode effect of "normal" diodes. What also kills p-i-n diode effect is ca[acitance and this is where schottkey' come in as they have "high" capacitance over the junction. Zener diodes also do quite a good job in reducing p-i-n diode effects. These are the effects that can be measured and is quoted as EMI.
See article http://ixdev.ixys.com/DataSheet/4592a0fc-7250-493a-a3e4-c5c1c69dc02e.pdf
Vishay makes a fast soft recovery diode which I think is worth a try 20ETF... it is 20A rated and insulated.
Hope my response added some value.
Remember to sit back and listen to music and not only listen for faults ;-)
I would suggest using discrete schottkey's and making your own bridge. You don't need bypass caps when using schottkey unless you want an overkill design to be sure.
Try these diodes from Vishay http://www.vishay.com/docs/94093/94093.pdf
They are specifically designed for low EMI which is what this whole fast diode bridge theory is about.
Those Vishay are 10A. I want the diode bridge to be 20A or more.
These Onsemi might fit the bill:
http://www.onsemi.com/pub_link/Collateral/MBR40250-D.PDF
I need to find out who's selling them.
Anyone tried them?
These Onsemi might fit the bill:
http://www.onsemi.com/pub_link/Collateral/MBR40250-D.PDF
I need to find out who's selling them.
Anyone tried them?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.