Bridge Rectifier vs. Schottky Diodes for F4

[KevinHeem]

This is the last item I need to add to my parts order for building an F4, so I need to make a decision between using the universal PSU with Schottky diodes or a bridge rectifier block.

I built my F5 using the diyAudio Universal PSU with Schottky Diodes.

I'm leaning towards the rectifier block as it saves space in a 4U chassis, and the easy of installation.

I just wonder if I'm giving anything up?

 

[Nelson Pass]

It's debatable. Use whatever makes you happy.

 

[KevinHeem]

Unfortunately life is generally not about making me happy...in fact it's usually about making someone else happy. But your right, my hobby should make me happy.

I'm going with the bridge rectifier block. It will make me happy during installation.

 

[mcandmar]

Difference is about $40 vs $4. .....but i went with the Schottkys for my F4

 

[Nelson Pass]

The standard "best" we have at Pass Labs is the fast/soft recovery types,
but snubbing is the best thing whatever you use.

 

[KevinHeem]

but snubbing is the best thing whatever you use.

I've heard that before. But calculating the values seems to be over my head. I have read two articles...struggling with the math.

I guess I rely too much on the online calculators where you plug in some values and get an answer.

I'm using the diyAudio store power supply board and there is a spot for input snubbing.

 

[Mark Johnson]

You can build a bellringer test jig on your protoboard which will let you visually dial in the optimum value of the snubber. Search the forums for "bellringer" and/or "Quasimodo" -- you will find that the thread NAME is No-Math Snubbers which sounds well-suited to your preferences.

The article I like best is the one in Linear Audio magazine. Its takeaway was

48 different diodes were tested, including Schottkys, HexFreds, SoftRecovery, Silicon Carbide, you name it.

The very best performing diode, without a snubber, was not as good as the very worst performing diode plus a snubber
​

BTW the worst performing diode was a 35 amp, 1000 volt, silicon bridge rectifier. Part number GBPC3510. sales link

 

[Russellc]

Unfortunately life is generally not about making me happy...in fact it's usually about making someone else happy. But your right, my hobby should make me happy.

I'm going with the bridge rectifier block. It will make me happy during installation.

My first Pass, the F5 build, I used Peter Daniels rectifier boards and nice separate diodes, 16 in all. Lots of solder points. Insulator pads under each one. For the next one, I used blocks. Have used blocks on every Pass power amp since. So much easier, less space used, and so on. Close enough to make me happy!

I will check out these snubber articles however!

Russellc

 

[Bare]

Hmmm as above... built my F6 and PSU as per guides etc.
Been working beyond expectations for ~1 year now.
Now I read about snubbers and their 'goodness' .
My PSU's feature diode block bridges rather than discrete Fast & Slow diodes
Should I retrofit the rascals to my DIY shoppe pcb in the required places or is this just gilding the lily?

Somewhere in memory.. I read that Only 1 fast /slow diode is actually necessary in a bridge for benefit... Accurate?

 

[Mark Johnson]

I bet there's a good chance you could find out the answer, by running LTSPICE simulations with different numbers of excellent/poor diodes.

 

[mattworld]

Hi,

You can find here a good paper about snubbing diode :

http://www.hagtech.com/pdf/snubber.pdf

 

[Bare]

I bet there's a good chance you could find out the answer, by running LTSPICE simulations with different numbers of excellent/poor diodes.

Thanks.
Unfortunately I have no cabalilities in electro wizardy per se.
Let alone sim pgms.
Simple man looking for simple answers.
My simple question being: Will filling in the 'spaces' in the Diy Audio PS pcb.. make audible difference ?

 

[Mark Johnson]

Try it and see. Listen without, then listen with. If it makes a difference you are happy. If it doesn't make a difference, you've only spent a piddling $2.00 on extra, needless, parts. And in return for that money spent, you possess new knowledge.

 

[2 picoDumbs]

The standard "best" we have at Pass Labs is the fast/soft recovery types,
but snubbing is the best thing whatever you use.

Could you comment whether the parasitic capacitance in the twisted wires running too and from the bridge rectifiers adds anything in the way of snubbing?

 

[2 picoDumbs]

Try it and see. Listen without, then listen with. If it makes a difference you are happy. If it doesn't make a difference, you've only spent a piddling $2.00 on extra, needless, parts. And in return for that money spent, you possess new knowledge.

Is it possible just to caveman the cap value or is there a narrow window for the desired effect.
Eg if a diode requires 100nF and you add 470nF (approximately 4 to 5 times higher) will the higher value cap still add something beneficial?

 

[analog_sa]

The very best performing diode, without a snubber, was not as good as the very worst performing diode plus a snubber

The interesting question is whether "diode sound" equates to a different level and character of ringing. Somehow i suspect there is more to it.

 

[analog_sa]

Eg if a diode requires 100nF and you add 470nF (approximately 4 to 5 times higher) will the higher value cap still add something beneficial?

I have wondered the same. Is overdamping the resonance a good idea?

 

[Mark Johnson]

Have a look at the Quasimodo thread here on DIYA. It shows adjustable snubbers made with 20-turn trimmer potentiometers, allowing you (yes YOU) to slooooowly dial in an optimum snubber while watching the snubber do its job, in real time, directly on the face of your oscilloscope. The thread includes large numbers of scope photos, showing Before, During, and After the snubber optimization procedure.

What may astonish you, or make you angry, is that the Quasimodo thread absolutely refuses to connect anything across the rectifier diode(s). There are no caps in parallel with the diode(s). There are no RC series networks in parallel with the diode(s). Instead, the Quasimodo thread uses Linear Systems Theory to arrive at a surprising (to some people) conclusion: the right place for a snubber is across the transformer secondary. This may astonish you, or make you angry, or both. But the measured data, before and after, is right here on DIYA, if you want to have a look.

Let me remind you of something you already know: there are MANY well-loved and highly praised pieces of audio equipment, which don't contain any snubbers at all. Maybe they would sound even better if snubbers were added; or, maybe they would sound worse if snubbers were added. I sure haven't seen any threads where somebody writes "Hey I added snubbers to my Levinson ML-2 Class A power amp and the sound quality got much better!" So don't get overexcited about snubbers; it is possible to build excellent gear without them. Nelson Pass's 20 watt Class A amplifier project called "A20" has no snubbers, for example; yet many people absolutely love its sound. If you happen to own an A20, please pop in a snubber and find out whether the sound gets better, gets worse, or doesn't change. Then tell people what you discovered!

 

[jameshillj]

For people that play around with different types of components and variations of 'sound result', it's a bit of a surprise when you experience people that genuinely can't actually hear any difference in transformers, diodes, caps, resistors, etc in power supplies when listening to speakers but don't have any difficulty in discerning changes in components when listening to headphones on the same equipment
Unfortunately, many of these still insist that it doesn't matter as they listen only to speakers and totally resist the idea that any change in a power supply could deliver a better sound and continue using very average components

I find it's difficult to change preconceived attitudes when it comes to hearing/listening (I'm sometimes subject to this myself, unfortunately) and the term 'just try it for yourself' somehow seems to produce extraordinary resistance

 

[alazira]

The interesting question is whether "diode sound" equates to a different level and character of ringing. Somehow i suspect there is more to it.

For people that play around with different types of components and variations of 'sound result', it's a bit of a surprise when you experience people that genuinely can't actually hear any difference in transformers, diodes, caps, resistors, etc in power supplies when listening to speakers but don't have any difficulty in discerning changes in components when listening to headphones on the same equipment
Unfortunately, many of these still insist that it doesn't matter as they listen only to speakers and totally resist the idea that any change in a power supply could deliver a better sound and continue using very average components

I find it's difficult to change preconceived attitudes when it comes to hearing/listening (I'm sometimes subject to this myself, unfortunately) and the term 'just try it for yourself' somehow seems to produce extraordinary resistance

Mark Johnson's quasimodo is a true gem for the DIY community. This is one of the few slam dunk (Jordan 360-behind-the-back) improvements you can do. Even better, you can get into the game for under a Benjamin (cheapomodo kit + inexpensive oscilloscope) and a little reading.

IMO, however, although you can eliminate diode ringing in even the poorest parts, this does not make different diodes sound the same. So if you are inclined to voice parts, I encourage you to do so.