Rectifier diodes with bypass cap´s & gounding speaker basket´s

[mikaelmark]

You´re findings are interesting, as I also intend to use the snubber for my Lehmann Clone but I´m a bit confused as the recommended values for snubbers are 47nF - 1uF and 4,7Ohm - 47Ohm. But however, if you have measured, I will trust you!

https://diyaudioheaven.wordpress.com/tutorials/power-supplies/snubbering/

Did you leave the four 0,1uF MKT or did you removed them before measuring? Those bypass caps should leak some AC, and this is the reason I want to replace them with snubber instead. I think they should not be used together!

Also, do you recommending me to leave the PC-filter for the primary side together with the snubber?

And what kind of resistor should be best; Metal film, Carbon dioxide, Carbon composition or MOX? Some people says the Carbon composition (CCR) will be best here, due the risk of high Wattage spikes for snubbers.

 

[betwixt]

Based on what we have already said, Carbon dioxide resistors would be best but you might have trouble soldering to them.

Really, the snubber circuits are not going to make one iota of difference to the audio performance but if you insist on using them, any kind of resistor except wire-wound will work equally well and the capacitor type is unimportant too. If you use 0.1uF capacitors the AC leakage will be about 600uA peak, probably less than 0.01% of the rest of the circuit.

Brian.

 

[prairiedog]

Audio electronics design is full of superstition and speculation. There might be some voodoo involved. I just use a scope or spectrum analyzer and measure things.

The snubber values depend on the transformer's power, voltage and construction (toroid vs EI).
I think the DIY-Audio Heaven values are out to lunch by a factor of 10. 1uF is just silly huge and 4.7 ohms is low unless a SMPS. 10-100nF and 10-1,000ohms I've used, at 1/4W for under ~40V and 1-3W for high voltages 85-600VAC. I gave values for EI transformer typical for a DAC or headphone amp. The switching spikes are high peak value but low energy so MOX or carbon comp are fine.

I find any capacitor directly across the transformer secondary or across the rectifier diodes shifts the ringing down to the 100's kHz. I'd removed the four caps to do the tests.

Rectifier switching RFI is significant when scaled up for audio power amplifiers. Big diodes, 500-2000VA power transformers - then the RFI is stronger. Prove it with an EMI probe or AM radio on the line cord or near the amplifier's wiring, or a look-see with a scope.
So I add snubbers for low level audio and metrology circuits, just to keep RFI to a minimum, as good practice.

A mains filter would interact such a tiny amount with a secondary snubber, it's not a factor. Leave it in.

 

[c_mitra]

Really, the snubber circuits are not going to make one iota of difference to the audio performance...

True. As long as it works. A good power supply can make several iota:s of difference. A simple capacitor snubber across the rectifier diodes can reduce noise from the mains (power line noise) and also reduces spikes from unknown sources.

True that the noise will be filtered by regulators and the filter caps- both before and after the regulators- will take up most of it. But the voltage spikes can kill the diodes slowly and definitely. And they will behave unpredictably just before they die.

Fast diodes have lower internal capacitance and they are more vulnerable to voltage spikes. Using a capacitor in parallel with a fast diode is counter productive (beyond by comprehension).

Line voltage is quite noisy (not as clean as you think) for most of the people in the world. And the capacitors are not expensive. They do help (IMHO).

 

[mikaelmark]

You´re findings are interesting, as I also intend to use the snubber for my Lehmann Clone but I´m a bit confused as the recommended values for snubbers are 47nF - 1uF and 4,7Ohm - 47Ohm. But however, if you have measured, I will trust you!

https://diyaudioheaven.wordpress.com/tutorials/power-supplies/snubbering/

Did you leave the four 0,1uF MKT or did you removed them before measuring? Those bypass caps should leak some AC, and this is the reason I want to replace them with snubber instead. I think they should not be used together!

Also, do you recommending me to leave the PC-filter for the primary side together with the snubber?

 

[c_mitra]

but I´m a bit confused as the recommended values for snubbers are 47nF - 1uF and 4,7Ohm - 47Ohm...

You have given a link and I have glanced through that.

The problem with "their" argument is that switching produces ringing (noise) which is true.

But a regular diode switches at zero crossing (without a big filter capacitor on the output side).

Regular diodes are slow and that reduces switching noise (even with a big filter capacitor)

For general use, take a small value capacitor (0.1 uF) and a small value resistor (2-10 Ohm).

It will also filter part of the noise coming from the line side (power line) - you cannot assume that the regulator is going to take care of everything.

It is good to take care of the turn off noise close to the source.

 

[KlausST]

Hi,

Regular diodes are slow and that reduces switching noise

I don´t agree. Slo diodes may increase switching noise.
Many (all?) Diodes usually are fast at turn ON. From high impedance to low impedance.
But slow diodes are slow at turn OFF.
Slow diodes have high reverse recovery time. They don´t become high impedance when they should ... they stay conductive with negative current. Then - after some time - they release which causes the noise.

Fast diodes switch OFF more fast, more close to zero current. Thus they reduce noise.

Klaus

 

[mikaelmark]

Re: Rectifier diodes with bypass cap´s & gounding speaker basket´s

Sorry for my double posting, I thought my first post did´nt went through one week ago and thank´s really much for all your recommendations!

Prairiedog; Will you leave the your snubber in the amp or will you swap back to the original four 0,1uF caps? And is it ok for you if I refer to your posts at this forum at the head-fi forum about this amp, as I found them really useful?

https://www.head-fi.org/threads/the...hmann-black-cube-linear-clone.501046/page-108

- - - Updated - - -

Audio electronics design is full of superstition and speculation. There might be some voodoo involved. I just use a scope or spectrum analyzer and measure things.

The snubber values depend on the transformer's power, voltage and construction (toroid vs EI).
I think the DIY-Audio Heaven values are out to lunch by a factor of 10. 1uF is just silly huge and 4.7 ohms is low unless a SMPS. 10-100nF and 10-1,000ohms I've used, at 1/4W for under ~40V and 1-3W for high voltages 85-600VAC. I gave values for EI transformer typical for a DAC or headphone amp. The switching spikes are high peak value but low energy so MOX or carbon comp are fine.

I find any capacitor directly across the transformer secondary or across the rectifier diodes shifts the ringing down to the 100's kHz. I'd removed the four caps to do the tests.

Rectifier switching RFI is significant when scaled up for audio power amplifiers. Big diodes, 500-2000VA power transformers - then the RFI is stronger. Prove it with an EMI probe or AM radio on the line cord or near the amplifier's wiring, or a look-see with a scope.
So I add snubbers for low level audio and metrology circuits, just to keep RFI to a minimum, as good practice.

A mains filter would interact such a tiny amount with a secondary snubber, it's not a factor. Leave it in.

I know many of you say any resistor should be fine, but some say carbon will do better here. Why? Metal film usually do better with lower noise etc while carbon is a older model more used in former equipment!

So except the power ratings, who of those below may be best suited for this snubber (I´m not sure if this is carbon composite or carbon dioxide)?

**broken link removed**

**broken link removed**

 

[KlausST]

Hi,

some say carbon will do better here. Why?

The bigger noise compared to metal film resistors here is of no interest. Any other noise around this resistor is magnitudes higher.

But carbon resistors have better pulse rating compared to metal film resistors.

Consider: a 10 Ohms snubber resistor needs to widhtstand full sine peak voltage across it for a short time.
At 230V RMS the peak voltage is about 325V. connected to an uncharged capacitor ... this means there will be about 32A of short peak current flow causing about 10,000W of short peak power.
With a 100nF capacitor this current drops to about one third of current each microsecond (tau= R x C) .. causing the power to be reduced to about 1/10 each microsecond.
10,000W --> after 1 us --> 1,350W --> after 1 us --> 180W --> after 1 us --> 24W --> after 1 us --> 3.3W --> after 1 us --> 0.45W

Note: These are theoretical values for zero line impedance.

Klaus

 

[mikaelmark]

Hi,


The bigger noise compared to metal film resistors here is of no interest. Any other noise around this resistor is magnitudes higher.

But carbon resistors have better pulse rating compared to metal film resistors.

Consider: a 10 Ohms snubber resistor needs to widhtstand full sine peak voltage across it for a short time.
At 230V RMS the peak voltage is about 325V. connected to an uncharged capacitor ... this means there will be about 32A of short peak current flow causing about 10,000W of short peak power.
With a 100nF capacitor this current drops to about one third of current each microsecond (tau= R x C) .. causing the power to be reduced to about 1/10 each microsecond.
10,000W --> after 1 us --> 1,350W --> after 1 us --> 180W --> after 1 us --> 24W --> after 1 us --> 3.3W --> after 1 us --> 0.45W

Note: These are theoretical values for zero line impedance.

Klaus

Thank´s, KlausST!

Are Carbon the same as Carbon composite and/or Carbon dioxide?

Do you think this will be correct?: **broken link removed**

 

[KlausST]

Hi,

Do you think this will be correct?

All we can do is guess.

If you want a reliable answer then buy a resistor from a reliable source with detailed informations like resistor_maufacturer and exact type or partnumber.

Klaus

 

[c_mitra]

Are Carbon the same as Carbon composite and/or Carbon dioxide?

There is nothing like a carbon dioxide resistor. Two common types are composite and film.

For low values (say less than 1M) you can usually ignore the thermal noise. There are other types of noises that are dependent on the manufacturing process. Usually film resistors have lower noise.

There are metal film resistors that may have still lower noise figures but your application does not need it.

For your application, a 10 Ohm resistor (say about 2W; even 1W will do) may be sufficient.

You see that the resistor type is not at all critical. It will dissipate less than 0.5W in the normal operation.

 

[mikaelmark]

There is nothing like a carbon dioxide resistor. Two common types are composite and film.

For low values (say less than 1M) you can usually ignore the thermal noise. There are other types of noises that are dependent on the manufacturing process. Usually film resistors have lower noise.

There are metal film resistors that may have still lower noise figures but your application does not need it.

For your application, a 10 Ohm resistor (say about 2W; even 1W will do) may be sufficient.

You see that the resistor type is not at all critical. It will dissipate less than 0.5W in the normal operation.

Thank´s, and sorry if I´m nagging!

Before I will purchase some resistors, the one I´m thinking of if Carbon film.

Is film good choise as snubber?

And what do you meant with "Nothing is like Carbon dioxide", -is film a subcategory to composition or are there three different kind of Carbons?

 

[c_mitra]

is film a subcategory to composition or are there three different kind of Carbons?

Carbon has basically two allotropes: graphite and diamond. Of these only graphite is conducting.

Ignore some of the nano forms of carbon- nanotubes, fullerenes etc etc.

Carbon dioxide is a gas with a chemical formula of CO2- it does not have any use as a resistor. It is sometimes used in lasers.

Composition resistors are basically carbon powder (graphite) mixed with a binder and the proportion of carbon determines the resistance. Film resistors are basically thin deposits of carbon or metals on a ceramic substrate.

Metal film resistors have low temp coefficient. Nichrome is the common alloy widely used in metal film resistors.

The capacitor stores energy (max) 0.5x0.1x10^-6x310^2 per cycle; this energy must be dissipated in the resistor. Per second this becomes 0.1W

I just recommended 1W just to worry about the peak values.

 

[mikaelmark]

I will now go for prairedog´s recommended MKS 02/K2, 33nF with a carbon film 330 Ohm, but are not quite sure what voltage rating the cap should have; should 63 VDC be enough for this snubber?

https://www.wima.de/en/our-product-range/capacitors-in-pcm-2-5/

Merry Christmas to all of you!

 

[betwixt]

The capacitor has to be rated at a minimum of twice the peak AC voltage from the transformer. The diode itself will prevent the voltage on one half cycle from rising to more than about 0.6V (Vf of the diode) but on the other half cycle, when the diode isn't conducting, it will have the opposite polarity peak voltage on one side and the previously rectified DC voltage on the other.
Peak AC voltage is 1.414 times the RMS voltage so the rating must be more than 2.828 times the transformer voltage but to cater for transients and to add a safety margin, aim for about 50% higher voltage rating.

Example: 15V RMS = 21.25V peak, capacitor must withstand 2 * 21.25 = 42.5V and add 50% = 63.75V.
If you relax the 50% safety margin a little, yes you can use 63V rated capacitors.

Merry Christmas, or Nadolig LLawen as we say in my country.

Brian.

 

[mikaelmark]

As I also gonna use this Snubber for my DAC that has an R-Core transformer, with a total of five cables for secondary as follosing: 9-9, 15-0-15 VAC (one cable from each winding are connected together for 0), I want to know if the Snubbers should either be between 15 and 0 and another Snubber between 0 and 15 for the other winding.

The above will require a total of three Snubbers (inclusive one for the 9 VAC winding).

Or should it be only two Snubbers, so one Snubber will go between 15-15 (and not to the middle 0) and the other Snubber for 9 VAC?

 

[betwixt]

What is an "R-Core" transformer?

We need to backtrack here - regardless of the advice that you don't need them, are you proposing snubbers across the AC connection from the transformer or across each rectifier diode?

Brian.

 

[mikaelmark]

It´s similar to an toroidal transformer, but it´s winded a bit different and not like a donut as a toroid.



All I can tell is that the result of using a Snubber with a 47nF MKT in serial with a 330Ohm between the secondary for my headphone amp works excellent! But now I also want to replace the existing parallel caps for my WM8741 DAC, in a similar way. But as I has told, I´m not sure if there should be only one Snubber for the two 15VAC windings OR two Snubbers going from each 15VAC to 0VAC (other side of the windings, connected together)?

 

[betwixt]

That is just a normal toroidal transformer with primary and secondary on different bobbins. That construction is normally used when very high voltage isolation is needed between primary and secondary but in a domestic application it has little if any advantage and is more expensive to manufacture. They do have slightly lower capacitive coupling between windings but the shield in a conventional toroid does pretty much the same thing.

It makes very little difference which way you wire the snubbers, both the configurations will do the same thing.

Brian.