Jump to content

Do you snub your rectifiers?


Recommended Posts

I’ve got this amp recapping project on the cooker (where it will be for some time, because I’m a procrastinator), and while planning the replacement of caps I thought of other improvements I could make while I have the amp, a Rotel RB-991, in bits.

 

At some point I got onto the topic of rectifier diodes – how the fast, soft recovery diodes are preferred etc. The Rotel uses the run-of-the-mill BR154 rectifier, an easy target for upgrades I thought. I came across a paper by Mark Johnson in Linear Audio where he investigates how soft recovery diodes improve the ringing in conjunction with the transformer’s secondary winding. A paper worth reading, IMHO, with lots of measurements for dozens of diode types. But what caught my eye was a remark in the conclusion that soft recovery diodes can reduce ringing big time, but a common CRC network or even just snubbing the diodes with a suitable capacitor can remove it altogether. I also learned that tight twisting of the DC supply cables will help this situation by maximising capacitance between the load and the rectifier. I filed that away under “interesting”.

 

Then, I happened across a schematic of the Rotel RB-971 amp, the little brother of mine, and lo and behold – it uses 10nF snubbing caps across the rectifier diodes (BR84)! No idea why Rotel gave up on this practice, fitting 8 extra caps must have been a huge effort.

 

Long story short, I’m definitely trying this potential upgrade on my amp, but I was wondering whether this was something people consider doing with their DIY power amp projects? Especially those amps with unregulated power rails, like the classic Rotel A/B amps. Are there any cons, or just pros?

 

 

 

Edited by Steffen
oops
Link to post
Share on other sites
2 hours ago, Steffen said:

I’ve got this amp recapping project on the cooker (where it will be for some time, because I’m a procrastinator), and while planning the replacement of caps I thought of other improvements I could make while I have the amp, a Rotel RB-991, in bits.

 

At some point I got onto the topic of rectifier diodes – how the fast, soft recovery diodes are preferred etc. The Rotel uses the run-of-the-mill BR154 rectifier, an easy target for upgrades I thought. I came across a paper by Mark Johnson in Linear Audio where he investigates how soft recovery diodes improve the ringing in conjunction with the transformer’s secondary winding. A paper worth reading, IMHO, with lots of measurements for dozens of diode types. But what caught my eye was a remark in the conclusion that soft recovery diodes can reduce ringing big time, but a common CRC network or even just snubbing the diodes with a suitable capacitor can remove it altogether. I also learned that tight twisting of the DC supply cables will help this situation by maximising inductance and capacitance between the load and the rectifier. I filed that away under “interesting”.

 

Then, I happened across a schematic of the Rotel RB-971 amp, the little brother of mine, and lo and behold – it uses 10nF snubbing caps across the rectifier diodes (BR84)! No idea why Rotel gave up on this practice, fitting 8 extra caps must have been a huge effort.

 

Long story short, I’m definitely trying this potential upgrade on my amp, but I was wondering whether this was something people consider doing with their DIY power amp projects? Especially those amps with unregulated power rails, like the classic Rotel A/B amps. Are there any cons, or just pros?

 

 

Very interesting, S ... but I think you meant "tight twisting of the DC supply cables will help this situation by minimising inductance - and thus maximising capacitance - between the load and the rectifier"?

 

Inductance & capacitance act in opposite directions - ie.:

  • lower inductance comes with higher capacitance, and
  • lower capacitance comes with higher inductance.

Andy

 

  • Like 1
Link to post
Share on other sites

Tried that on several projects - no obvious "improvements" but nice thing to have on standrad bridge rectifiers or 1NXXXX silicon diodes. That being said, using them on modern fast diodes like MUR2030  or MUR1060 does not make a lot of sense. Those I usueally leave by themselves.

 

A correct snubbing requires a resisitor as well and generally serves to attenuates HF ringing of the diode recovery current. Caps alone help a bit but not much. It is easier just to have fast recovery diodes. Some info: 

 

https://electronics.stackexchange.com/questions/408119/how-do-you-calculate-rc-snubber-values-for-a-discrete-bridge-rectifier

  • Like 2
Link to post
Share on other sites
41 minutes ago, andyr said:

 

Very interesting, S ... but I think you meant "tight twisting of the DC supply cables will help this situation by minimising inductance - and thus maximising capacitance - between the load and the rectifier"?

 

 

Yes sorry, I’ll fix that.  After forever praising my own super-low inductance and fairly high capacitance speaker cable, you’d think I get that right... :) 

Link to post
Share on other sites


15 minutes ago, Decky said:

A correct snubbing requires a resisitor as well and generally serves to attenuates HF ringing of the diode recovery current. Caps alone help a bit but not much. It is easier just to have fast recovery diodes. Some info: 

 

 

Yep, Johnson also goes into how to dimension those. You only need one RC across the secondary.

 

The one cap for each diode thing probably works slightly differently, and is also seen occasionally (e.g. in that Rotel amp). It might be considered as easier, because even though you need more caps, the dimensioning seems to be rather non-critical (~4.7-47nF foil, or something like that). For a proper RC (or CRC) snubber you’d need to know the inductance of the secondary winding.

 

It looks like the Linear Audio article are behind a paywall, so here’s a link to the article I downloaded from somewhere a while ago. Not great quality, and a couple of pages are upside-down, but interesting nevertheless:

 

https://www.icloud.com/iclouddrive/0DuhUhpwDT5SXV6XXiT8YO3RA#Linear_Audio_-_Soft_Recovery_Diodes_Lower_Transformer_Ringing_by_10-20X 

Link to post
Share on other sites

Does all this technical sales pitch in audiophile electronics engineering explain why standard run of the mill power supply diodes are the most reliable and stable components for AC to DC Conversion, I’m talking about standard full bridge rectifiers used for 230V 10A 60/50hz switching.   The ringing on the on/off crossover is at least stable at zero volt and doesn’t move when temp increases, and one thing for sure doesn’t inherent reverse leakage that can be destructive to electrolytics.   I would stick to the standard diode, if your worried about diode switching ringing, which should be the last on your agenda to resolve as it has bugger all effect compare to introducing fast soft recovery diodes that has extremely worrying reverse leakage. 

Edited by Addicted to music
Link to post
Share on other sites

 

How does this "ringing" between the diodes and the transformer secondary, get past the usual DC smoothing and filtering network that follows the rectification in every power supply?

Link to post
Share on other sites
2 minutes ago, aussievintage said:

 

How does this "ringing" between the diodes and the transformer secondary, get past the usual DC smoothing and filtering network that follows the rectification in every power supply?

The assumption that’s been made is the usual BS that Ringing is HF  can leak into signal path, to snub it correctly you need to take into consideration of the winding of the transformer.   

Link to post
Share on other sites


1 hour ago, Addicted to music said:

The assumption that’s been made is the usual BS that Ringing is HF  can leak into signal path,

 

Well proper construction and layout should prevent that.  After all, people also worry about noise coming down the mains supply into the equipment, so the same precautions deal with all the above.

  • Like 1
Link to post
Share on other sites

Here quoted are Rod Elliot's thoughts on the subject of snubbers in d.c. power supplies - (source is Linear Power Supply Design (sound-au.com)

 

"Another 'interesting' myth is that using a snubber (essentially a Zobel network consisting of a series resistor and capacitor) across the diodes or transformer secondary windings will "improve the sound".  It won't do anything of the sort, but in some cases may reduce conducted EMI (electro-magnetic interference conducted back to the mains via the power lead) if this happens to be a problem.  This is covered in detail in the article Snubbers For Power Supplies - Are They Necessary And Why Might I Need One?.  As with many of the other mythical/ magic 'ingredients' that can be added to power supplies, they won't hurt anything, but they also will not change the 'sound' of the DC.  By all means include snubbers if it makes you feel better, but don't go posting nonsense on-line about the "amazing difference" they make.  It doesn't happen unless your internal wiring is poorly laid out and you get buzz as a result of internal radiated magnetic fields."

 

As with all things audio, the schools of thought are many and varied; therefore everyone may have different ideas, opinions and experiences.  As a consequence I don't always dismiss an idea simply because I disagree with it initially.  However I do appreciate it when opinions or ideas are backed up with some repeatable or verifiable evidence beyond the just "because I said so" anecdotes. *

 

* the rationale being that IME very few people actually openly admit to blowing thousands of dollars on an idea or product that reaped little or no "improvement" beyond increasing their credit card debit balance.  Not that I'm suggesting that adding snubbers across the transformer secondary winding(s) or across rectifier diodes is going to amount to a significant monetary investment. If you reap a worthwhile improvement then by all means implement it. ? ?

 

Cheers,

Alan R. 

 

  

 

 

 

  • Like 1
Link to post
Share on other sites
4 hours ago, aussievintage said:

 

How does this "ringing" between the diodes and the transformer secondary, get past the usual DC smoothing and filtering network that follows the rectification in every power supply?

Because it adds as noise carried,  then by the same DC circuit.  Think of diodes  and their reverse bias leakage  as a  semiconductor form of noise, akin to reactance as adding after the moment like a decay.   in photography terms like an image with insufficient pixels to make the image clear.   

 

Far better is to use thyristors as they offer complete rectification,  where the current rating drawn by a given circuit,  remains well below the devices rating. For thyristors their ratings are massive, so are very unlikely ever to be approached. When a thyristors rating is approached it begins to also create  a reverse bias leakage on one of its junctions  however does so slowly. With ratings available up to 3kA  audio purpose is easily catered for. The other big advantage is the thyristor is a controlled device, vs a diode which is uncontrolled.  

 

The puzzle of trying to design for when the thyristor triggers can be solved by using small conventional bridge rectifiers to power the gate of each thyristor . Here is such a circuit. Current draw in this circuit is just 3ma, the thyristors in TO92 package,  have current rating to 800ma each.

 

 

IMG_1705.JPG

Link to post
Share on other sites

Nah, I don't snub my rectifiers.  Not usually.   I have Mark Johnsons Quasimodo test jig and have measured many transformer/rectifier combinations and the process works a charm according to what the oscilloscope shows.

 

The only transformers I do snub are those attached to 'service' components such as fans and power protocols because I don't mind increasing the impedance of the power supply for those circuits and because I don't want those transformers ringing back into the mains line.  Have used lots of soft recovery schottky's in more important circuits in my builds, but am now shifting to active rectifiers primarily due to low heat generation (almost zero voltage drop) whilst being at least equivalent to the schottky's in terms of noise.  Don't know if they sound any better...don't really care because they seem to do the job, even in the high voltage valve circuits. 

 

Might even replace some schottky's with the active rectifiers in some of my existing builds just to cut the heat down a little and to free up some space. 

Edited by acg
Link to post
Share on other sites


It isn't trivial to measure the impact of doing it.

 

... and that it the crux of the problem.

 

Because many if not most people "can't" .....   and "in theory" that it would be good thing to do...... the only way to assess it is to "listen it it".

 

... and of course it sounds better.   ;) 

Link to post
Share on other sites

Ah Steffen, you're looking at the well known Hagerman snubber network for transformer secondary windings - a capacitor in parallel to a capacitor + series resistor network - for these purposes it's 10nF in // to 150nF + a  measured resistance, yes?

 

Circuit 'snubbers' have been around for a long time and extensively used in power engineering and it works very well indeed as 'acg' mentioned above, to suppress/dampen the ringing in the secondary windings of the transformer produced by the 'reverse charge/back firing 'spike' of the conventional diodes and especially the classic BP block bridges.

 

 The big question is "does it make a difference to the sound" and the reason for the varied results is determined by the circuit that the power supply drives - the mainly simple classA circuits (gain stages, filters, etc) that I tend to use are quite dependent on the power supply for the overall sound and the snubber filters are clearly a benefit but other circuits that have a designed rejection to power supply 'noise' are nearly immune, hence the wide difference in opinions - as it's pretty simple and inexpensive to add the components, why not just try it and see for yourself - the resistor is the variable, not the caps, hence the Quazimodo Bell Ringer test unit.  

 

 I have no idea why 'snake-oil' is mentioned here as it's basic engineering.

 

These synchronous fet bridge rectifiers are marvelous devices - the fets show less than 10mR series resistance so passing 3+ amps thru them produces milliwatts heat and so surface mount small pcbs is just fine and no heatsinks, etc plus a much 'smoother' sound for my use.  I haven't tried them in class AB amplifiers or shunt reg power supplies yet so can't add anything there 

 

They're widely used in the auto industry 

 

 

Link to post
Share on other sites
19 hours ago, HdB said:

Ah Steffen, you're looking at the well known Hagerman snubber network for transformer secondary windings - a capacitor in parallel to a capacitor + series resistor network - for these purposes it's 10nF in // to 150nF + a  measured resistance, yes?

 

Circuit 'snubbers' have been around for a long time and extensively used in power engineering and it works very well indeed as 'acg' mentioned above, to suppress/dampen the ringing in the secondary windings of the transformer produced by the 'reverse charge/back firing 'spike' of the conventional diodes and especially the classic BP block bridges.

 

 The big question is "does it make a difference to the sound" and the reason for the varied results is determined by the circuit that the power supply drives - the mainly simple classA circuits (gain stages, filters, etc) that I tend to use are quite dependent on the power supply for the overall sound and the snubber filters are clearly a benefit but other circuits that have a designed rejection to power supply 'noise' are nearly immune, hence the wide difference in opinions - as it's pretty simple and inexpensive to add the components, why not just try it and see for yourself - the resistor is the variable, not the caps, hence the Quazimodo Bell Ringer test unit.  

 

 I have no idea why 'snake-oil' is mentioned here as it's basic engineering.

 

These synchronous fet bridge rectifiers are marvelous devices - the fets show less than 10mR series resistance so passing 3+ amps thru them produces milliwatts heat and so surface mount small pcbs is just fine and no heatsinks, etc plus a much 'smoother' sound for my use.  I haven't tried them in class AB amplifiers or shunt reg power supplies yet so can't add anything there 

 

They're widely used in the auto industry 

 

 

Could it be better not to use a band aid solution, rather deal with eliminating reverse charge/back firing spike of conventional diodes, by using a better device ?  

Link to post
Share on other sites

Yes, Chris, I wondered that myself and did a round of trying a whole lot of different types of diodes, the 3 layer Crees, etc even those giant 'hockey-puck' ones but ended up using the older 'more conventional Soft Recovery' types and the Hagerman snubbers cancelled the ringing and this suited my purpose quite well for the circuits I use. 

Some years ago, we developed a system of 'free wheeling' zero crossing relays using triacs and these could be adapted to high power rectifiers like the Swanns but couldn't get rid of residual noise - maybe this is more in your area with the thyristor control systems

 

I have found that the synchronous rectifiers give a simpler and better result to any systems I've previously used but they're quite a bit more expensive even using 10mR dual fets - those 'tvi' guys over in Romania (?) use something like 3mR devices but unknown where they come from - these are regarded as perhaps 'the best' bridges available but again, not at all cheap - they built a couple for me to use with a rather 'u-beaut' EI centre-tapped transformer but still awaiting absurd postal delivery times  over 3 months! 

 

 As you know, the type of caps used in the power supplies is probably more significant so I can see the hesitation about "going down the rabbit hole" of diode choice - probably a matter of choosing priorities even if the $s aren't a major factor. 

Link to post
Share on other sites


  • Recently Browsing   0 members

    No registered users viewing this page.




×
×
  • Create New...