Re: Choosing Amplifier Power Diodes

[gene ohara]

Tim,

Thanks for your suggestions!

It has recently been recommended to me to use a 35A chassis mounted
bridge rectifier with 400V rating (for amps <500W) to minimize
destruction due to line spikes, surges.... Does everyone agree?

Am still confused about what normally destroys the power diodes: 1.
the destruction of the output semiconductor, 2. a AC line power surge,
or 3. other?

Thanks again everyone!!

Gene




Browntimdc > wrote in message >...
> (gene ohara) wrote in
> om:
>
> > Am beginning to design a non switching audio amp power supply. Would
> > like to make it to be able to withstand power surges or if damaged
> > have it confined to inexpensive components so am trying to research
> > what components are important and what can be generic.
> >
> > Question: For non switching power supplies have power diodes reached
> > generic status or are there differences (all the same size)?
> >
> > Appreciate any suggestions, esp by someone who has already done all
> > this.
> >
> > Thanks
> >
> > Gene
>
> For 50/60 Hz PS circuits generic components are fine. You'll want them
> rated about 3X your average current to make up for the high RMS/average
> current ratio (more heating) because the diodes conduct in fairly narrow
> pulses. They'll be fine for surges. It's a good idea also to put an RC
> snubber (~0.01uF in series with 33 ohm) across each diode to kill ringing
> and the resulting RF noise.
>
> A certain denizen who used to hang around rec.audio would claim you need
> ultra fast diodes, but that's just not the case.
>
> I've been doing this for 30 years.
>
> Tim

[GregS]

In article >, wrote:
>On Wed, 09 Jul 2003 19:37:52 GMT, Browntimdc
> wrote:
>
(gene ohara) wrote in
om:
>>
>>> Am beginning to design a non switching audio amp power supply. Would
>>> like to make it to be able to withstand power surges or if damaged
>>> have it confined to inexpensive components so am trying to research
>>> what components are important and what can be generic.

Isn't that what a surge supressor strip is for?
I hate the term surge but it can happen. I'm mostly concerned with spikes.
Switching supplies are the ones which burn out. They don't have a transformer
in there to buffer the AC line.

>>
>>For 50/60 Hz PS circuits generic components are fine. You'll want them
>>rated about 3X your average current to make up for the high RMS/average
>>current ratio (more heating) because the diodes conduct in fairly narrow
>>pulses. They'll be fine for surges. It's a good idea also to put an RC
>>snubber (~0.01uF in series with 33 ohm) across each diode to kill ringing
>>and the resulting RF noise.

A typical 100 WPC amplifier might draw 300 or more watts total. Thats
maybe 4 amps at max power. Average may be less than 1 amp. Each diode only
works 1/2 cycle, so thats 1/2 amps average current per diode. Peaks can
normally ok through diodes, but I know through experiance, the bridges frequently
go bad, mostly by shorting, and the fuse will blow.

>
>You don't need a network unless you have diodes in series, such as
>that found in high voltage power supplies. The circuit impedance is
>too low for ringing and this wouldn't work on rf noise anyway. All
>this circuit will do is make the diodes appear leaky, very leaky.

[Rich Andrews]

(gene ohara) wrote in
m:

> Tim,
>
> Thanks for your suggestions!
>
> It has recently been recommended to me to use a 35A chassis mounted
> bridge rectifier with 400V rating (for amps <500W) to minimize
> destruction due to line spikes, surges.... Does everyone agree?
>
> Am still confused about what normally destroys the power diodes: 1.
> the destruction of the output semiconductor, 2. a AC line power surge,
> or 3. other?
>
> Thanks again everyone!!
>
> Gene
>
>


In my experience failure is caused most by one factor. Too low a PIV rating
for the application. When replacing diodes, I always used 1500 or more PIV
rated diodes with a current rating of 2.5 A or more. Never had a piece of
equipment come back for PS problems afterwards. The originals diodes were
300 volt PIV 1 amp diodes in many cases. In larger pieces I used 3 amp 1500
PIV diodes. In power amps I used the highest PIV and current rated bridge
rectifiers I could find that would fit well. I did have an issue with one
piece of gear that everytime it rained, the PS regulation transistor would
blow. A MOV across the line solved that problem permanently.

Rarely I have seen line surges. Many line spikes, but very few surges. I
define surges as an overvoltage condition that last for 1 second or more.
Moderate to severe line spikes/transients can be 5000 Volts or more.

Good luck.

r



>
>
> Browntimdc > wrote in message
> >...
>> (gene ohara) wrote in
>> om:
>>
>> > Am beginning to design a non switching audio amp power supply.
>> > Would like to make it to be able to withstand power surges or if
>> > damaged have it confined to inexpensive components so am trying to
>> > research what components are important and what can be generic.
>> >
>> > Question: For non switching power supplies have power diodes reached
>> > generic status or are there differences (all the same size)?
>> >
>> > Appreciate any suggestions, esp by someone who has already done all
>> > this.
>> >
>> > Thanks
>> >
>> > Gene
>>
>> For 50/60 Hz PS circuits generic components are fine. You'll want them
>> rated about 3X your average current to make up for the high RMS/average
>> current ratio (more heating) because the diodes conduct in fairly
>> narrow pulses. They'll be fine for surges. It's a good idea also to put
>> an RC snubber (~0.01uF in series with 33 ohm) across each diode to kill
>> ringing and the resulting RF noise.
>>
>> A certain denizen who used to hang around rec.audio would claim you
>> need ultra fast diodes, but that's just not the case.
>>
>> I've been doing this for 30 years.
>>
>> Tim
>



--
"Any sufficiently advanced technology is indistinguishable from
magic."

Arthur C. Clarke (1917 - ), "Technology and the Future"

[Pete KE9OA]

Soft knee point diodes are a good way to go...............these types don't
radiate all of that high energy RFI the way that other types do. If you
don't opt for this type of diode, there are ways around it. If you connect
..01uF caps across each rectifier diode, this will prevent them from
rectifying RF. Another necessary thing to do is connect high voltage 1uF
film caps from each leg of the secondary of the power transformer to ground.
This knock out the hash below 500kHz. At 120Hz, 1uF provides a high enough
reactance, so you don't have much current flowing through them. For higher
voltage supplies that operate at +/- 50 volts or higher, use .22 to .33uF.
Just make sure that you use caps with a 500V or higher rating. I made this
discovery, when I started building VLF receivers.

Pete

Rich Andrews > wrote in message
.44...
> (gene ohara) wrote in
> m:
>
> > Tim,
> >
> > Thanks for your suggestions!
> >
> > It has recently been recommended to me to use a 35A chassis mounted
> > bridge rectifier with 400V rating (for amps <500W) to minimize
> > destruction due to line spikes, surges.... Does everyone agree?
> >
> > Am still confused about what normally destroys the power diodes: 1.
> > the destruction of the output semiconductor, 2. a AC line power surge,
> > or 3. other?
> >
> > Thanks again everyone!!
> >
> > Gene
> >
> >
>
>
> In my experience failure is caused most by one factor. Too low a PIV
rating
> for the application. When replacing diodes, I always used 1500 or more
PIV
> rated diodes with a current rating of 2.5 A or more. Never had a piece of
> equipment come back for PS problems afterwards. The originals diodes were
> 300 volt PIV 1 amp diodes in many cases. In larger pieces I used 3 amp
1500
> PIV diodes. In power amps I used the highest PIV and current rated bridge
> rectifiers I could find that would fit well. I did have an issue with
one
> piece of gear that everytime it rained, the PS regulation transistor would
> blow. A MOV across the line solved that problem permanently.
>
> Rarely I have seen line surges. Many line spikes, but very few surges. I
> define surges as an overvoltage condition that last for 1 second or more.
> Moderate to severe line spikes/transients can be 5000 Volts or more.
>
> Good luck.
>
> r
>
>
>
> >
> >
> > Browntimdc > wrote in message
> > >...
> >> (gene ohara) wrote in
> >> om:
> >>
> >> > Am beginning to design a non switching audio amp power supply.
> >> > Would like to make it to be able to withstand power surges or if
> >> > damaged have it confined to inexpensive components so am trying to
> >> > research what components are important and what can be generic.
> >> >
> >> > Question: For non switching power supplies have power diodes reached
> >> > generic status or are there differences (all the same size)?
> >> >
> >> > Appreciate any suggestions, esp by someone who has already done all
> >> > this.
> >> >
> >> > Thanks
> >> >
> >> > Gene
> >>
> >> For 50/60 Hz PS circuits generic components are fine. You'll want them
> >> rated about 3X your average current to make up for the high RMS/average
> >> current ratio (more heating) because the diodes conduct in fairly
> >> narrow pulses. They'll be fine for surges. It's a good idea also to put
> >> an RC snubber (~0.01uF in series with 33 ohm) across each diode to kill
> >> ringing and the resulting RF noise.
> >>
> >> A certain denizen who used to hang around rec.audio would claim you
> >> need ultra fast diodes, but that's just not the case.
> >>
> >> I've been doing this for 30 years.
> >>
> >> Tim
> >
>
>
>
> --
> "Any sufficiently advanced technology is indistinguishable from
> magic."
>
> Arthur C. Clarke (1917 - ), "Technology and the Future"
>

[gene ohara]

Rich and Pete

Thanks for the detailed comments!!

What diode mfr(s) do you recommend?

Gene



"Pete KE9OA" > wrote in message >...
> Soft knee point diodes are a good way to go...............these types don't
> radiate all of that high energy RFI the way that other types do. If you
> don't opt for this type of diode, there are ways around it. If you connect
> .01uF caps across each rectifier diode, this will prevent them from
> rectifying RF. Another necessary thing to do is connect high voltage 1uF
> film caps from each leg of the secondary of the power transformer to ground.
> This knock out the hash below 500kHz. At 120Hz, 1uF provides a high enough
> reactance, so you don't have much current flowing through them. For higher
> voltage supplies that operate at +/- 50 volts or higher, use .22 to .33uF.
> Just make sure that you use caps with a 500V or higher rating. I made this
> discovery, when I started building VLF receivers.
>
> Pete
>
> Rich Andrews > wrote in message
> .44...
> > (gene ohara) wrote in
> > m:
> >
> > > Tim,
> > >
> > > Thanks for your suggestions!
> > >
> > > It has recently been recommended to me to use a 35A chassis mounted
> > > bridge rectifier with 400V rating (for amps <500W) to minimize
> > > destruction due to line spikes, surges.... Does everyone agree?
> > >
> > > Am still confused about what normally destroys the power diodes: 1.
> > > the destruction of the output semiconductor, 2. a AC line power surge,
> > > or 3. other?
> > >
> > > Thanks again everyone!!
> > >
> > > Gene
> > >
> > >
> >
> >
> > In my experience failure is caused most by one factor. Too low a PIV
> rating
> > for the application. When replacing diodes, I always used 1500 or more
> PIV
> > rated diodes with a current rating of 2.5 A or more. Never had a piece of
> > equipment come back for PS problems afterwards. The originals diodes were
> > 300 volt PIV 1 amp diodes in many cases. In larger pieces I used 3 amp
> 1500
> > PIV diodes. In power amps I used the highest PIV and current rated bridge
> > rectifiers I could find that would fit well. I did have an issue with
> one
> > piece of gear that everytime it rained, the PS regulation transistor would
> > blow. A MOV across the line solved that problem permanently.
> >
> > Rarely I have seen line surges. Many line spikes, but very few surges. I
> > define surges as an overvoltage condition that last for 1 second or more.
> > Moderate to severe line spikes/transients can be 5000 Volts or more.
> >
> > Good luck.
> >
> > r
> >
> >
> >
> > >
> > >
> > > Browntimdc > wrote in message
> > > >...
> > >> (gene ohara) wrote in
> > >> om:
> > >>
> > >> > Am beginning to design a non switching audio amp power supply.
> > >> > Would like to make it to be able to withstand power surges or if
> > >> > damaged have it confined to inexpensive components so am trying to
> > >> > research what components are important and what can be generic.
> > >> >
> > >> > Question: For non switching power supplies have power diodes reached
> > >> > generic status or are there differences (all the same size)?
> > >> >
> > >> > Appreciate any suggestions, esp by someone who has already done all
> > >> > this.
> > >> >
> > >> > Thanks
> > >> >
> > >> > Gene
> > >>
> > >> For 50/60 Hz PS circuits generic components are fine. You'll want them
> > >> rated about 3X your average current to make up for the high RMS/average
> > >> current ratio (more heating) because the diodes conduct in fairly
> > >> narrow pulses. They'll be fine for surges. It's a good idea also to put
> > >> an RC snubber (~0.01uF in series with 33 ohm) across each diode to kill
> > >> ringing and the resulting RF noise.
> > >>
> > >> A certain denizen who used to hang around rec.audio would claim you
> > >> need ultra fast diodes, but that's just not the case.
> > >>
> > >> I've been doing this for 30 years.
> > >>
> > >> Tim
> > >
> >
> >
> >
> > --
> > "Any sufficiently advanced technology is indistinguishable from
> > magic."
> >
> > Arthur C. Clarke (1917 - ), "Technology and the Future"
> >

[Rich Andrews]

(gene ohara) wrote in
om:

> Rich and Pete
>
> Thanks for the detailed comments!!
>
> What diode mfr(s) do you recommend?
>
> Gene
>
>

International Rectifier comes to mind.

r


>
> "Pete KE9OA" > wrote in message
> >...
>> Soft knee point diodes are a good way to go...............these types
>> don't radiate all of that high energy RFI the way that other types do.
>> If you don't opt for this type of diode, there are ways around it. If
>> you connect .01uF caps across each rectifier diode, this will prevent
>> them from rectifying RF. Another necessary thing to do is connect high
>> voltage 1uF film caps from each leg of the secondary of the power
>> transformer to ground. This knock out the hash below 500kHz. At 120Hz,
>> 1uF provides a high enough reactance, so you don't have much current
>> flowing through them. For higher voltage supplies that operate at +/-
>> 50 volts or higher, use .22 to .33uF. Just make sure that you use caps
>> with a 500V or higher rating. I made this discovery, when I started
>> building VLF receivers.
>>
>> Pete
>>
>> Rich Andrews > wrote in message
>> .44...
>> > (gene ohara) wrote in
>> > m:
>> >
>> > > Tim,
>> > >
>> > > Thanks for your suggestions!
>> > >
>> > > It has recently been recommended to me to use a 35A chassis mounted
>> > > bridge rectifier with 400V rating (for amps <500W) to minimize
>> > > destruction due to line spikes, surges.... Does everyone agree?
>> > >
>> > > Am still confused about what normally destroys the power diodes:
>> > > 1. the destruction of the output semiconductor, 2. a AC line power
>> > > surge, or 3. other?
>> > >
>> > > Thanks again everyone!!
>> > >
>> > > Gene
>> > >
>> > >
>> >
>> >
>> > In my experience failure is caused most by one factor. Too low a PIV
>> rating
>> > for the application. When replacing diodes, I always used 1500 or
>> > more
>> PIV
>> > rated diodes with a current rating of 2.5 A or more. Never had a
>> > piece of equipment come back for PS problems afterwards. The
>> > originals diodes were 300 volt PIV 1 amp diodes in many cases. In
>> > larger pieces I used 3 amp
>> 1500
>> > PIV diodes. In power amps I used the highest PIV and current rated
>> > bridge rectifiers I could find that would fit well. I did have an
>> > issue with
>> one
>> > piece of gear that everytime it rained, the PS regulation transistor
>> > would blow. A MOV across the line solved that problem permanently.
>> >
>> > Rarely I have seen line surges. Many line spikes, but very few
>> > surges. I define surges as an overvoltage condition that last for 1
>> > second or more. Moderate to severe line spikes/transients can be 5000
>> > Volts or more.
>> >
>> > Good luck.
>> >
>> > r
>> >
>> >
>> >
>> > >
>> > >
>> > > Browntimdc > wrote in message
>> > > >...
>> > >> (gene ohara) wrote in
>> > >> om:
>> > >>
>> > >> > Am beginning to design a non switching audio amp power supply.
>> > >> > Would like to make it to be able to withstand power surges or if
>> > >> > damaged have it confined to inexpensive components so am trying
>> > >> > to research what components are important and what can be
>> > >> > generic.
>> > >> >
>> > >> > Question: For non switching power supplies have power diodes
>> > >> > reached generic status or are there differences (all the same
>> > >> > size)?
>> > >> >
>> > >> > Appreciate any suggestions, esp by someone who has already done
>> > >> > all this.
>> > >> >
>> > >> > Thanks
>> > >> >
>> > >> > Gene
>> > >>
>> > >> For 50/60 Hz PS circuits generic components are fine. You'll want
>> > >> them rated about 3X your average current to make up for the high
>> > >> RMS/average current ratio (more heating) because the diodes
>> > >> conduct in fairly narrow pulses. They'll be fine for surges. It's
>> > >> a good idea also to put an RC snubber (~0.01uF in series with 33
>> > >> ohm) across each diode to kill ringing and the resulting RF noise.
>> > >>
>> > >> A certain denizen who used to hang around rec.audio would claim
>> > >> you need ultra fast diodes, but that's just not the case.
>> > >>
>> > >> I've been doing this for 30 years.
>> > >>
>> > >> Tim
>> > >
>> >
>> >
>> >
>> > --
>> > "Any sufficiently advanced technology is indistinguishable from
>> > magic."
>> >
>> > Arthur C. Clarke (1917 - ), "Technology and the Future"
>> >
>



--
"Any sufficiently advanced technology is indistinguishable from
magic."

Arthur C. Clarke (1917 - ), "Technology and the Future"

[Trevor Wilson]

"gene ohara" > wrote in message
m...
> Tim,
>
> Thanks for your suggestions!
>
> It has recently been recommended to me to use a 35A chassis mounted
> bridge rectifier with 400V rating (for amps <500W) to minimize
> destruction due to line spikes, surges.... Does everyone agree?
>
> Am still confused about what normally destroys the power diodes: 1.
> the destruction of the output semiconductor, 2. a AC line power surge,
> or 3. other?

**Yes to all the above. Additionally, switch on surge is a factor in diode
destruction. At switch on, the filter capacitors appear as a short circuit.
If the power transformer is of sufficiently low impedance and the capacitors
large enough and of very low internal resistance, then if the amp is
switched on at the top of the sine wave, a large enough peak current may
flow to damage the diodes. Some equipment employs surge limit circuits to
prevent this damage.


--
Trevor Wilson
www.rageaudio.com

[Bruce Burke]

On Mon, 14 Jul 2003 19:57:27 GMT, "Trevor Wilson"
> wrote:

>
>"gene ohara" > wrote in message
m...
>> Tim,
>>
>> Thanks for your suggestions!
>>
>> It has recently been recommended to me to use a 35A chassis mounted
>> bridge rectifier with 400V rating (for amps <500W) to minimize
>> destruction due to line spikes, surges.... Does everyone agree?
>>
>> Am still confused about what normally destroys the power diodes: 1.
>> the destruction of the output semiconductor, 2. a AC line power surge,
>> or 3. other?
>
>**Yes to all the above. Additionally, switch on surge is a factor in diode
>destruction. At switch on, the filter capacitors appear as a short circuit.
>If the power transformer is of sufficiently low impedance and the capacitors
>large enough and of very low internal resistance, then if the amp is
>switched on at the top of the sine wave, a large enough peak current may
>flow to damage the diodes. Some equipment employs surge limit circuits to
>prevent this damage.
>
>
>--
>Trevor Wilson
>www.rageaudio.com
>
>
Yes: IF the power transformer......

Most have a high enough internal impedance and phase shift that they
will limit the turn-on surge. Otherwise, you'd have to replace your
on-off switch all the time.


Regards,

Bruce
Hitting reply is futile, use the following:
).

[Nothing40]

TCS > wrote in message ol.net>...
> <html><input type crash></html>
> begin
> On 9 Jul 2003 13:24:43 -0700, gene ohara > wrote:
> > Kevin McMurtrie
> >
> > Thanks for the tips!
> >
> > Do you have any suggestions or links for sizing power doides ... Am
> > thinking that would prefer to destroy the power diodes rather than
> > the more expensive output devices if there was a power surge or other
> > prob. Do you agree?
>
> No. That's what fuses are for.

In my experience,by the time the fuse blows,its already too late for the output's.

[TCS]

On 19 Jul 2003 06:46:45 -0700, Nothing40 <f> wrote:
>TCS > wrote in message ol.net>...
>> <html><input type crash></html>
>> begin
>> On 9 Jul 2003 13:24:43 -0700, gene ohara > wrote:
>> > Kevin McMurtrie
>> >
>> > Thanks for the tips!
>> >
>> > Do you have any suggestions or links for sizing power doides ... Am
>> > thinking that would prefer to destroy the power diodes rather than
>> > the more expensive output devices if there was a power surge or other
>> > prob. Do you agree?
>>
>> No. That's what fuses are for.

>In my experience,by the time the fuse blows,its already too late for the output's.

I agree.

The best way to have equipment last, and that includes speakers, is never to
run them into distortion. When raising the volume makes the music "sound
fuller" instead of "sound louder", it's too high.