| youyoung21147 |
Hello,
I'm planning to build a PSU for my future chip amp (surely a GainClone) and I am wondering which rectifiers I should use.
I plan to order some samples from OnSemi to compare but I ask here what kind of rectifier I should use :
I know they need to be fast, 35ns is a good value.
But concerning the rest I don't know. Why do people like MUR860 diodes more than the other ones ?
I was thinking that 8A for 2 amplifiers isn't enough.
So why not use 16A or 30A ? Are small ones more "refined" ? Or is bigger better ?
Concerning the rating, should it be adapted to the circuit or is it better to use higher voltage ratings ? (eg. 200V or 600V)
Finally are there any other parameters to take into consideration when a choice has to be made ???
I'm a little lost so every comment is welcome :) |
|
|
| mzzj |
| quote: | Originally posted by youyoung21147
Hello,
I know they need to be fast, 35ns is a good value.
But concerning the rest I don't know. Why do people like MUR860 diodes more than the other ones ?
|
Not only fast, but also soft-recovery.
---
They are in fashion :D
-----
Why not use schotky-diodes. ? Faster than ultrafast and absolutely soft-recovery. plus low-losses. OK, if you need more than 100V rated diodes then schotkies are not first choice. (altough 200v ones are becoming more availlable and 600&1200V SiC-schotkies are also availlable.) |
|
|
| wintermute |
The MUR860's are 50ns but have very high (600V) ratings... go to www.farnell.com and do searches on soft recovery and ultra fast recovery, you should get a large selection :) some as fast as 20ns whilst still handling the 8A of the MUR860's..... not sure on the voltage.....pobably 200V max for that speed...
as for which ones work well???????
Tony. |
|
|
| bezobraznia |
Check BYW29E-150, BYW29E-200, BYV-series etc.
Right now I'm with first type on my Gainclone and with similar for my DeathOfZen-Class-A amp and I'm very satisfied with results.
Great diodes - ultra fast, fast recovery etc.
Best Reggards
Niki / Bulgaria |
|
|
| planet10 |
| quote: | Originally posted by mzzj
also soft-recovery. |
That is actually the important bit.
| quote: | | Why not use schotky-diodes. ? Faster than ultrafast and absolutely soft-recovery. plus low-losses |
Any time your required voltage is low enough these are hard to beat...
dave |
|
|
| janneman |
| quote: | Originally posted by bezobraznia
[snip]- ultra fast, fast recovery etc.
Best Reggards
Niki / Bulgaria |
... which is the same of course. Ultra-fast refers to the recovery, NOT to the forward path.
Jan Didden |
|
|
| youyoung21147 |
Mmmm Interesting !
: Not only fast, but also soft-recovery.
==> How do you quantify "softness" of diodes ?
Is it worthy to use some 16A or 30A for more "safety" ?
Is it better to use 600V instead of 200V ? |
|
|
| bezobraznia |
| 10x for correction... but You follow my point :) |
|
|
| youyoung21147 |
I found some "Ultrasoft" rectifiers at OnSemi : the MSR1560.
Rated 600V 15A
They are said to be ultrasoft, and have a pretty good 45ns Trr.
I assume they are good candidates aren't they ?
I also found some Shottky rectifiers : MBR40250.
They have a good 35ns Trr and are rated 40A 250V
They are oversized, but I assume it doesn't matter.
Apart from price, what model would you recommend me between those ?
Are there measurable parameters apart from Trr that would help me guess which or which part is the best ?? |
|
|
| Sch3mat1c |
I'm missing something, what exactly is it about 8.33mS pulses that needs switching 10^5 times faster?
Tim |
|
|
| janneman |
| quote: | Originally posted by Sch3mat1c
I'm missing something, what exactly is it about 8.33mS pulses that needs switching 10^5 times faster?
Tim |
Tim,
The Trr is REVERSE recovery time, it has no bearing on the forward conductance. Diodes that have no fast recovery take some time to turn of after the voltage across them reverses, this is the recovery time. The recovery often behaves oscillatory and that generates hf pulsels that can get into the supply. However, when you have a fast recovery, you can get one single sharp turn-off pulse which is also bad. Unless they are soft recovery: smooth, fast turn off is best.
This is the theory, and in smps applications, when the signal to be rectified is 100's of kHz or even a Mhz, this is measurable. Some people apparently feel they can HEAR it in audio as well, one more piece of audio folklore...
Jan Didden |
|
|
| forr |
I think Nelson Pass recomands the use of standard diodes.
~~~~~~~ Forr
§§§ |
|
|
| jean-paul |
| quote: | Originally posted by janneman
Some people apparently feel they can HEAR it in audio as well, one more piece of audio folklore...
Jan Didden |
Jan, did you never do the simple comparison tests ? Even I hear differences in power supply diodes. Differences, no more no less. I feel it is slippery to contribute the differences to the speed of the diode though.... |
|
|
| Peter Daniel |
| quote: | Originally posted by forr
I think Nelson Pass recomands the use of standard diodes.
|
Not anymore ;)
| quote: | Originally posted by Nelson Pass
We are currently using the ST Micro BYW99W-200, which is
a dual "high efficiency, fast recovery" 35 amp diode at 200
volts, and have been very happy with it.
It's remarkable for us to have converted to exotic rectifiers,
as we previously were very sceptical, but we did find a distinct
sonic improvement, and they also made it easier to meet
RF noise emission standards. :cool: |
http://www.diyaudio.com/forums/show...1830#post321830
And Mr. Carr suggests that "sonic differences are caused not only by the category of rectifier diode, but the manufacturer as well"
http://www.diyaudio.com/forums/show...2326#post102326 |
|
|
| bezobraznia |
I think that using fast diodes put some requirement on all capacitance. Better do similar connection like following sheme example:
transformer -> ultrafast diodes bridge -> capacitance 24000uF -> capacitance 7500uF -> capacitance -> 2200uF -> amplifier itself
Reggards
Niki
EDIT: hehe, Peter Daniel like i say before - using of BYW or BYV fast Philips diodes are great advantage for sheme itself! ;) |
|
|
| Nordic |
Hi I'm looking through my rummage bins for diodes to rectify 6Vac 1A supply, to drive a .1A fan and some LEDs...
Only ones I can find 4 of is FR104...
I looked up the datasheet but it confuses me...
It says maximum reverse voltage 400V but maximum forward voltage 1V, does this men one cn only use these diodes on 1Vac? |
|
|
| analog_sa |
| quote: | | does this men one cn only use these diodes on 1Vac? |
Of course not. It means that the forward biased drop will not exceed 1v at max current. |
|
|
| janneman |
| quote: | Originally posted by analog_sa
Of course not. It means that the forward biased drop will not exceed 1v at max current. |
... or, said in another way, they need up to 1V across them before they start to conduct into a load that draws an amp...
Jan Didden |
|
|
| janneman |
| quote: | Originally posted by janneman
... or, said in another way, they need up to 1V across them before they start to conduct into a load that draws an amp...
Jan Didden |
... or said in yet another way, the dynamic forward resistance may be up to an ohm.. ;)
Jan Didden |
|
|
| Nordic |
Yep, I noticed they dropped the power quite a bit (well by a large margin considering the total value of this small winding)...
Don't know if its them or the fan, but it is makeing noise on the amp windings.... tried connecting a 100R and 100ufd cap in parallel with it.. no difirence.. |
|
|
| youyoung21147 |
| I've seen tutorials on PSU building and they seem to shunt the rectifier's pins with 22nf capacitors to short the oscillations, making the PSU "cleaner". I think I'll try this one day ! |
|
|
| wintermute |
This is a quite a standard practice for normal diodes, but is not recommended for soft recovery diodes, because it actually counteracts their soft recovery mechanism (or so I read) :)
Tony. |
|
|
| youyoung21147 |
Wow, nice to learn it :)
It makes a few parts less in the circuit, i.e less cost, less error risk...
I think I'll go for OnSemi 15A ultrasoft/ultrafast rectifiers, they seem to fit well my purpose ! |
|
|
| Christer |
This paper has been recommended before by others, so this is just a reminder for those who missed it. It is highly recommended reading about soft recovery, snap factors and everything else regarding reverse characteristics of diodes.
www.microsemi.com/micnotes/302.pdf
Edit:
Some more reading:
http://personales.ya.com/lcardaba/f...0high%20end.pdf
There are also a few pages about this in ON Semis "Rectifier Application Handbook" HB214, which unfortunately I couldn't find on their site right now (is it no longer available?). |
|
|
| wintermute |
| Thanks Christer :) |
|
|
| jneutron |
| quote: | Originally posted by Christer
This paper has been recommended before by others, so this is just a reminder for those who missed it. It is highly recommended reading about soft recovery, snap factors and everything else regarding reverse characteristics of diodes.
www.microsemi.com/micnotes/302.pdf
Edit:
Some more reading:
http://personales.ya.com/lcardaba/f...0high%20end.pdf
There are also a few pages about this in ON Semis "Rectifier Application Handbook" HB214, which unfortunately I couldn't find on their site right now (is it no longer available?). |
The microsemi paper is very good. The only thing I would take issue with is the oscillation above zero for the snap.. The paper kinda implies the diode does that, whereas it is actually the circuit's reaction to the abrupt change of the diode conduction that does it.
For the circuit limited Irm, I still laugh at the slope from If through the zero current line..back in '82, I put a Trr circuit together that did that in 250 pSec..it allowed the diffusion guys to see what the 15 nSec product under development was really doing.
As for that Jon Schleisner guy...why would anybody believe what he has to say??? Geeze, his employer only shipped 11 million diodes per day back in '93..(fun guy, worked with him..)
I like his transient suppressor writeups also..
What suprises me though, is I don't see anything talking about the conduction across the xfmr winds during recovery time, with the resultant PC currents..and how snubs can make that worse..
Cheers, John |
|
|
| KBK |
Anyone care to test diodes sonically? I know we all do, but, there is one aspect of diodes I'm particularly interested in. How about running a resistively loaded, DC signal, with AC signal overlay, and recover that signal? Analog throughout, of course. And put diodes in that output path. And keep switching them out for different ones. If you capacitively load your observation point*, it makes a great post to observe L, C, R and junction, separately or as a group. And check out your combinations, before you even build a circuit.
*(As in listen to, directly as possible: A headphone-centric circuit would be most valuble here) |
|
|
| janneman |
| quote: | Originally posted by KBK
Anyone care to test diodes sonically? I know we all do, [snip] |
No. I don't:D
Jan Didden |
|
|
| KBK |
What I meant was, that in the end, we all do, when we use them in a circuit. But attempting to separate the sound of the given actions of the diode is interesting. When conducting, or how the transient behaviour revelas itself in terms how it turns off and what exactly it does, effect wise. Etc, etc.
I've done crazier things than that - and each one led to an understanding of each given component or component 'effect', and allowed me to surmise and deduct and head to places in 'audiophile design understandings' most folk haven't been to yet. :) |
|
|
| janneman |
| quote: | Originally posted by KBK
What I meant was, that in the end, we all do, when we use them in a circuit. But attempting to separate the sound of the given actions of the diode is interesting. When conducting, or how the transient behaviour revelas itself in terms how it turns off and what exactly it does, effect wise. Etc, etc.[snip] |
No, I don't. I design my circuits in such a way as to prevent diode issues to make it to the speaker connectors.
Any amp that is showing sound differences with different diodes in a repeatable, blind test for me is an incompetent design...:cool:
Jan Didden |
|
|
| KBK |
| May I politely suggest, that you are missing the point...? :) |
|
|
| SY |
| quote: | Originally posted by janneman
No, I don't. I design my circuits in such a way as to prevent diode issues to make it to the speaker connectors.
Any amp that is showing sound differences with different diodes in a repeatable, blind test for me is an incompetent design...:cool:
Jan Didden |
Perfect.
One very direct measurement which is rarely done is spectral analysis of supply voltage under differing load conditions. In the limited amount I've done, a circuit with the grounding done right just doesn't show a whole lot of difference between exotic and conventional rectifier technology in 50/60 Hz sine wave circuits. |
|
|
| janneman |
| quote: | Originally posted by KBK
May I politely suggest, that you are missing the point...? :) |
Entirely possible, it's known to have happened before.;) .
In that case could you explain then what the point is, I have enough self-confidence, to believe I could understand your explanation.
Until now I thought "sonically testing rectifiers" meant to evaluate rectifiers by the changes they make to the sound of a particular amplifier?
Jan Didden |
|
|
| poobah |
KBK,
| quote: | | No, I don't. I design my circuits in such a way as to prevent diode issues to make it to the speaker connectors. |
Diodes, balanced AC, PSU transformers, all of these things could make a difference. The fact remains that if they do make a difference, it indicates poor power supply design. Consider the perfect diode in a PSU. It will actually generate more energy (spectra) at higher frequencies that a regular recovery diode, despite how "nasty" it appears on a scope. Is this good or bad? It depends on the total design.
SY's remark about spectral testing under dynamic conditions is right on the money.
The push for faster, softer, smoother diodes has come about in response to the needs of SMPS design where operation is in kHz & MHz... snubbers and switching locus are a huge part of all that as well.
:) |
|
|
| janneman |
| quote: | Originally posted by poobah
KBK,
Diodes, balanced AC, PSU transformers, all of these things could make a difference. The fact remains that if they do make a difference, it indicates poor power supply design. Consider the perfect diode in a PSU. It will actually generate more energy (spectra) at higher frequencies that a regular recovery diode, despite how "nasty" it appears on a scope. Is this good or bad? It depends on the total design.
SY's remark about spectral testing under dynamic conditions is right on the money.
The push for faster, softer, smoother diodes has come about in response to the needs of SMPS design where operation is in kHz & MHz... snubbers and switching locus are a huge part of all that as well.
:) |
Yes, it all makes a difference in electrical terms. Yet, if I had designed an amp that sounded audibly different with a different diode, I'd hide in a corner and tell no one about it...
Jan Didden |
|
|
| rdf |
| quote: | Originally posted by SY
In the limited amount I've done, a circuit with the grounding done right just doesn't show a whole lot of difference between exotic and conventional rectifier technology in 50/60 Hz sine wave circuits. |
Hi SY, do you recall how much was 'not a whole lot'? Actually I'm a bit surprised it was measurable. |
|
|
| john curl |
| I went to soft recovery rectifiers about 10 years ago, after we did an A-B test on a Vendetta power supply. It took me YEARS to actually believe that it could make a difference, but it did. I just upgraded Dave Wilson's Vendetta phono stage with soft recovery diodes. It is not easy to do in this case, as we did not assign much room in the power supply box, but the results are worth it. This is an example of how I continually attempt to update my older products, just to keep up with the latest efforts of others. I can now laugh at the reluctance of those who don't think that soft recovery diodes are necessary. Oh well, just more sonic wins for me. ;) |
|
|
| SY |
| quote: | Originally posted by rdf
Hi SY, do you recall how much was 'not a whole lot'? Actually I'm a bit surprised it was measurable. |
Differences are less than the repeatability of the measurements. It may indeed not be measurable, at least not with the 24/192 soundcard I use. That's why I'm somewhat skeptical- DC really is DC.
I used high speed diodes in my preamp's supply, just for the hell of it, since they were pretty cheap. And, IIRC, I used the same ones in the RLD power amp. No direct sonic comparisons so I don't know if there's any voodoo effect. |
|
|
| poobah |
| I could see the advantage in a phono stage. But doesn't it still point to filtering? |
|
|
| serengetiplains |
A PSU filter can be viewed as a voltage divider. In any voltage divider, some voltage goes one way (to ground, in a PSU), and some goes another (to your speakers, ultimately).
Here's a small thought experiment regarding the efficacy of power supply filters. Inject a large quantity of noise onto the AC line. Will *any* of that noise appear in a circuit fed by a power supply? Questions regarding the audibility of any noise coursing past the PSU filter can be laid to one side for this question. |
|
|
| Bgt |
| quote: | Originally posted by janneman if I had designed an amp that sounded audibly different with a different diode, I'd hide in a corner and tell no one about it...
Jan Didden |
Yep, I think so too, still cannot hear the difference with a normal setup with standard diodes or the schottky ones. |
|
|
| rdf |
Thx SY, that's what I figured. Not a whole lot is often a euphemism for squat measurable difference.
For what it's worth I haven't played with different silicon rectifiers but approach it by using a stage or two of filtering comprised of a uHenry-size RF inductor followed by a small film cap in the B+ supply to do the RF duties as most supply chokes resonate well into the audio band. I think it makes a worthwhile difference and as scientific proof offer burnt candles, feathers and deceased chickens: my personal voodoo. |
|
|
| poobah |
| quote: | | Inject a large quantity of noise onto the AC line. Will *any* of that noise appear in a circuit fed by a power supply? |
The math suggests a PSU would require infinite capacitance (with zero ESR/ESL) and resistance to achieve zero noise... audible or not. Also, a rather spendy method.
The point remains however, if your filter cannot reduce diode noise to acceptable levels... what other types of noise (line borne) are then welcome to pass right through?
RDF,
The chickens taste good afterwards as well.
:) |
|
|
| serengetiplains |
| Spendy indeed! Whether a given level of noise is acceptable depends on many factors, including one's experience. My acceptable level has consistently lowered over time. |
|
|
| rdf |
| There's no law stating all power supply impedances must be between B+ and ground. Spice sims of small air-core inductors between stages of capacitive filtering suggests huge gains in RF rejection. Cheap, cheerful and comes in BBQ or deep-fried. :) |
|
|
| john curl |
| How about series inductors AND soft recovery rectifiers? That's what we do. |
|
|
| serengetiplains |
| Those gains occur because you've changed the position of the tap on the voltage divider. Noise still divides between ground and amplifying device, etc. |
|
|
| serengetiplains |
| quote: | Originally posted by john curl
How about series inductors AND soft recovery rectifiers? That's what we do. |
Is what Meitner does also. Except I'm the guy who installs the rectifier. |
|
|
| poobah |
Yep... but these things can have decent numbers attached and some reduction in capacitor cost can be realized.
I think the diode quest falls apart when you ignore another PSU plugged into the same mains. For that matter, just look at the mains themselves!
RDF,
Play with some common-mode chokes someday... also ferrite beads... marvelous little friends they are.
I had to get a brush motor to pass CE testing once... in a REAL hurry. Looped the wires through a hex nut a few times and added a 0.1 uF cap... passed easily.
:) |
|
|
| rdf |
| quote: | Originally posted by john curl
How about series inductors AND soft recovery rectifiers? That's what we do. |
Sorry, no experience. I haven't progressed beyond thermonic diodes yet. =D |
|
|
| poobah |
Air cores are also much lower in saturated fat... healthy.
:) |
|
|
| KBK |
And for all you winos (SY) out there:
Get Thine A$$ down to the liquored-up store, wherever that maybe in your local locale, and buy the HECULA spanish wine that is making the rounds these days. Sells for $13.98 in Canada, (lower $in the states etc) and buy many a case. World ending numbers. Like it's the last bottle you will ever have. It scores a 92, on average, which puts it in the $50-75 a bottle price range, but it's only about $13 in the US. Supposedly, about 18,000 cases ended up being issued, so it should be easy to find. Matures late 2008, absolutely fine right now. |
|
|
| poobah |
| Whaddathink SY... is this a money maker? |
|
|
| KBK |
| yeah, you dip the diodes in this wine, yeah. Treated. Special. etc. With a side of low fat inductance. |
|
|
| SY |
| Well, the marketing campaign seems to be working, eh? |
|
|
| FastEddy |
see: http://en.wikipedia.org/wiki/Walter_H._Schottky / http://en.wikipedia.org/wiki/Schottky_diode
Inventor of the diode, really. And his first success was the best. All but forgotten until Shockly et al, invented the transisor ....
These diodes are ultra fast, quick recovery and almost magical when used in parallel with ordinary silicon diodes in producing a very fast recovery bridge rectifiers = "soft" recovery and in specialized detectors and microwave detectors ... and to prevent static discharge across CMOS chip inputs ... like op-amps.
... we couldn't have gigaHertz computers with out 'em.
:smash: |
|
|
| AndrewT |
Hi Fasteddy,| quote: | | These diodes are ultra fast, quick recovery and almost magical when used in parallel with ordinary silicon diodes in producing a very fast recovery bridge rectifiers |
that's new to me.
Can you elaborate? |
|
|
| FastEddy |
When used in parallel across the inputs of Complimentary Oxide Metal Semi conductor, Schottky type diodes protect these inputs from static, and dramatic or very rapid changes in voltage.
+
|
^
|
input
|
^
|
-
:smash: |
|
|
|
