It may sound odd, but this rating depends at least on: a) the type/manufacturer of the rectifier diode, b)pulse time. 
You should search some diode datasheets, because probably noone can say whether 50A rms diodes from a random manufacturer can handle 110A peak.
And there's another thing that you should consider. Are your pulses going to be repetitive (i.e. because of the loud music) or non-repetitive (i.e. because of capacitors' charging during power-on).
If second, you can use 30EPH06 diodes, that I've used in my PSU. They have have a continuous current rating of 30 A and a peak non-repetitive current of 300 A.
You should search some diode datasheets, because probably noone can say whether 50A rms diodes from a random manufacturer can handle 110A peak.
And there's another thing that you should consider. Are your pulses going to be repetitive (i.e. because of the loud music
) or non-repetitive (i.e. because of capacitors' charging during power-on).If second, you can use 30EPH06 diodes, that I've used in my PSU. They have have a continuous current rating of 30 A and a peak non-repetitive current of 300 A.
Here's some interesting diodes from Cree. They're silicon carbide schottky diodes and exhibit a positive tempco in their forward voltage, which means that you can parrallel them for higher current ratings, and with no reverse recovery time to speak of, and low leakge (at least for a schottky type diode), may be an interesting option:
http://www.cree.com/products/pdf/CSD20030.pdf
http://www.cree.com/products/pdf/CSD20030.pdf
No need for exotic, expensive diodes; if this is line frequency, standard recovery silicon diodes will suffice.
I can't think of a single diode made for power rectification that isn't rated for an absurd number of peak amperes. The 1N4001 is rated for 1A rms but 30A peak. I think the 35A FWBs I have on hand are rated for like 625A peak...
Unless you have an absurdly high input capacitor with absurdly low ESR, you don't have to worry about peak current at all. Besides, your transformer will melt from the massive RMS current long before the diodes will.
Tim
I can't think of a single diode made for power rectification that isn't rated for an absurd number of peak amperes. The 1N4001 is rated for 1A rms but 30A peak. I think the 35A FWBs I have on hand are rated for like 625A peak...
Unless you have an absurdly high input capacitor with absurdly low ESR, you don't have to worry about peak current at all. Besides, your transformer will melt from the massive RMS current long before the diodes will.
Tim
Heaven forbid some people would be interested in a little general knowledge to add to their repertoire.
IMHO, the whole point of DIY is lost if all you're doing is making stuff that the Pacific rim pumps out at a fraction of the cost that you can.
I once pulled some what appeared to be underated diodes out of the analog section in a Magnavox CD player and replaced them with some standard devices, thinking that it wouldn't effect the sound. To my surpriser it totally screwed up the sound. Had to put the originals back in. In my experience, the power supply very much effects the sound of a system, and even the lowly diodes can have an influence in a high definition system.
What, you actually listen?! And don't feel embarassed to reveal this to the experts?
More seriously, there are a couple of threads on the tweakers asylum about the Cree diodes. Exotic and expensive, maybe, but not to everyone's liking. I still haven't had a chance to plug in mine but if myth-busting is the theme of the day i might do it tonight.
I haven't tried them yet either. I've been waiting for the technology to mature a little, more specifically, I'm interested in the smaller one amp devices for pre-amp applications (if they ever make them available in a smaller package). As far as price goes, I would expect that to come down in time also, just as it has for what was once considered "exotic" blue LEDs, now commonplace.
I'd very much be interested in what conclusions you come to about the Crees when you try them.
It was the four diodes in the analog bridge. Very wimpy looking .5A switching diodes. As long as I was in there making some mods, I decided to replace them with some standard 1A rectifiers. I really didn't expect it to effect the sound, but man was I wrong. Looks like the designers of that unit knew something when they selected those diodes.
I agree, sometimes "current trends" make a difference and sometimes not. And when they do make a difference, the human psyche is such that it's often assumed that the difference must be for the better in all scenarios, which isn't necessarily the case. Historically speaking, observation have often preceded understanding, so to simply dismiss observations out of hand because it hasn't yet appeared in a text book isn't always forward looking. Taking too many claims seriously though is also a problem, as it can lead to misconceptions and wasted time. Where to draw the line seems to be the hardest part, as different people's experiences and backgrounds will effect that decision. This can of worms has been opened too many times in too many other places though. If you don't believe something has any effect on you, then don't bother with it personally. But the individual who started this thread clearly stated that he wanted fast recovery diodes for an application. If he states that is his desire, then that is his business, and in experimenting he may learn whether the diodes will actually make a difference in his application or not. An opinion is fine, but making nothing more then derogatory replys simply because you're of a different belief are not appropriate. I don't know whether diode type will make a difference in his particular instance or not, but I have seen cases where it does make a very significant difference, and others where it hasn't. IMHO the question of what the aspects are that effect whether the diodes make a minimal or major difference (whether it be related to PC board layout or whatever) is where the real story is.
hughmon said:
Nice post, agree to what you say. Indeed, it is difficult to separate the wheat from the chaff. Too many people avoid any intelligent discussion and just say "... but I heard a clear improvement. ". That of course kills all discussions, and I suspect that often it is just a made-up statement. That's anonymous internet for you, I guess.
Jan Didden
Schottky diodes switch very fast. That is an advantage for low loss if you have to switch high speed signals. In mains rectification that is no advantage at all. In fact, the fast switching is a disadvantage because the switch-off current cut-off generates lots of hf signal bursts that get into the supply and can muddle the audio. However, if you have a soft-recovery diode, then the hf is much less because the switch-off is very gradual.
So, either get a fast but soft recovery diode, or get a "normal" not too fast diode. Schottkys are too fast to be clean.
Jan Didden
So, either get a fast but soft recovery diode, or get a "normal" not too fast diode. Schottkys are too fast to be clean.
Jan Didden
Some/most "normal" diodes have indeed real nasty looking reverse recovery.janneman said:
But never heard that schottkies would be too fast, do you have some references about this? (no, i am not intrested to know how they sound like
)btw, when shall come new fashion about forward recovery? Even better as it is not usually specified on datasheets so somebody has to sacrifice his/hers ears for listening...
mzzj said:
Some/most "normal" diodes have indeed real nasty looking reverse recovery.
But never heard that schottkies would be too fast, do you have some references about this? (no, i am not intrested to know how they sound like
btw, when shall come new fashion about forward recovery? Even better as it is not usually specified on datasheets so somebody has to sacrifice his/hers ears for listening...
Well, it is all in the nature of the beast, on di/dt. If you switch (very) fast, you have a very large rate of change in current, the di/dt, the change in amperes per uSec. Like going from 1A conduction to zero in say 100nSec gives 10A/uSec di/dt (this is 10 million amps per second di/dt). Such a big di/dt gives a lot of hf bursts because of the wiring inductance and a few other things. (There is more involved like a kind of reverse current at switch off, but that quickly goes into a lot of detailed solid state physics). Standard electrical engineering. What soft recovery diodes do is to smoothen the change from full conduction to zero to minimise the hf bursts. Schottkys, by their fast nature, generate a lot of hf.
Jan Didden
Jan,
Maybe I have misunderstood something, but as I understand it, Schottky diodes are even better than any other soft-recovery diode, since they have almost no stored charge carriers, and thus almost no reverse current. See for instance this paper:
www.microsemi.com/micnotes/401.pdf
Maybe I have misunderstood something, but as I understand it, Schottky diodes are even better than any other soft-recovery diode, since they have almost no stored charge carriers, and thus almost no reverse current. See for instance this paper:
www.microsemi.com/micnotes/401.pdf
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.