Bypass mur860 diodes with caps?

[300_baud]

Hi there,

I was wondering if bypassing the mur860 diodes was a good idea. I've actually already tried this with 680 pf caps on a LM3886 gainclone. The sound seems a lot more relaxed than before. Can someone verify this? Is this a good thing or not?

Thanks for your help,
300_baud

 

[Nuuk]

Somebody told me recently that using these snubber caps on the rectifier bridge is good for higher frequencies and not so good for lower frequencies.

 

[300_baud]

So what values are people using to bypass their diodes? Like i mentioned in my first post, I'm using 680 pf.

Thanks again

 

[w00t]

I'm using 0.1uF X2 caps (Rifa PME271M). They are surplus from my Quietline project.

 

[300_baud]

Well, just wanted to report that the sound opened up, especially the treble. The transparency is about the same now as how it was without the caps on the diodes. Even though it is as transparent as before the caps, the treble doesn't feel as "sharp and as hard hitting" as it used to, which to me is a good thing. It's like the frequency response in the upper frequencies is more linear now.

Anyone else want to report their findings?

 

[Nuuk]

So what values are people using to bypass their diodes?

22n.

 

[Peter Daniel]

I had MSR860 in a DAC and there were there 0.1u orange caps across each diodes. I removed the caps later and the sound became much cleaner. I prefer it without caps.

But it may be circuit/system dependant, so do whatever you prefer.

 

[carlosfm]

Why do I always have a feeling that bypassing rectifier diodes is injecting AC on the PSU?

 

[Mark25]

Why do I always have a feeling that bypassing rectifier diodes is injecting AC on the PSU?

i was on the understanding that there's no current flow through a Capacitor, and the bypassing caps have quite high reactance to anything other than rf, so not much of the 50Hz "should" be getting through.
Mark

 

[head_spaz]

passing AC

is the AC operating at a higher frequency than the diode?

 

[Upupa Epops]

... is injecting AC to PSU ...

Yes, it is thruth. But both capacitors ( bypasing and PS ) make capacitive divider with rate 1 : 10 000 - 100 000 ( typicaly ), so we get attenuation of AC 70 - 80 dB. To this value we must add still 70 - 90 dB PSR of typical power amp and we get actual value of penetration AC ( caused by bypasing diodes ) to output signal : - 140 / - 170 dB. This value is minimal - 20 dB bellow noise floor.

 

[KT]

Not sure about the MUR860's, but I did try to bypass the FED-xxxx (can't rembember the numbers) diodes in my 47 Labs Power Humpty with .01uF Multicap RTX.

I expected an improvement, but the result was a sound so trebly and tilted up in the high ranges, that I removed the caps after a few days.

Don't know if the MUR860's are different, but it sure didn't work out on the Power Humpty diodes.

KT

 

[jackinnj]

Why do I always have a feeling that bypassing rectifier diodes is injecting AC on the PSU?

the diode's capacitance itself is in series with the inductance of the transformer secondary -- so you have a classical "tank" circuit -- the manufacturer's specification sheet will give you the capacitance at the voltage on the diode -- 100pF or thereabouts for a MUR860 at 30 volts.

An externally hosted image should be here but it was not working when we last tested it.

For the MBR20100, a Schottky Diode, the capacitance is much less, although ON-Semi doesn't have a graph like that above, a nominal value is given in their SPICE Model of 1x10e-11.

 

[analog_sa]

Why do I always have a feeling that bypassing rectifier diodes is injecting AC on the PSU?

Not just AC but lovely wide-bandwidth hash which otherwise will have to battle to pass through the diode's capacitances.

I wonder who first came with the half-baked idea to bypass diodes?

 

[jackinnj]

Not just AC but lovely wide-bandwidth hash which otherwise will have to battle to pass through the diode's capacitances.

I wonder who first came with the half-baked idea to bypass diodes?

It's pretty standard practice -- look at the why's and wherefore's on On-Semi's website -- or Fairchild Semi's website.

an incorrectly designed diode bridge will put out a nice several hundred millivolt RF signal at 600kHz -- about the bottom end of the a.m. radio band here in the states -- the high frequency ringing can be made worse by improper layout, etc.

the problem is worse for switching supplies.

 

[analog_sa]

an incorrectly designed diode bridge will put out a nice several hundred millivolt RF signal at 600kHz

Maybe using standard diodes but hardly with MUR860

 

[Elso Kwak]

Snubbers

Please have a look at this thread:
http://www.diyaudio.com/forums/showthread.php?s=&threadid=12539&highlight=high+speed+diode
The snubbers across the diodes were also discussed.

 

[jackinnj]

Maybe using standard diodes but hardly with MUR860

All right, I did the math for once, and I will do it again -- measure the secondary inductance with the primary leads shorted -- this should be ~100 to 200 uH. Measure the primary to secondary capacitance -- this could be a couple hundred picofarads. Obtain the diode capacitance from the manufacturer's data sheet -- vide above --

f= 1 / (2*pi* sqrt LC) = 700 kHz.

so, here's a snapshot of what the results are with and without a snubber, the diode is a MUR860, 220uF cap, 560R and 4.7Nf snubber:

An externally hosted image should be here but it was not working when we last tested it.

 

[samsagaz]

sorry for revive this old thread, but where i need to measure that kind of signal? where i need to place the osc probes?

 

[Gopher]

Not just AC but lovely wide-bandwidth hash which otherwise will have to battle to pass through the diode's capacitances.

I wonder who first came with the half-baked idea to bypass diodes?

Analog_SA I don't understand your logic here. Surely if you parallel a diode with a bypass capacitor you are increasing the total capacitance? Anyway, what you really want is low capacitance in the diode without a bypass to make the impedance to RFI on the mains very high.