Question about diode ratings in a rectifier bridge

[Ignoramus361]

Suppose that I build a rectifier bridge to rectify 60 Hz AC.

Each diode has a rating of Ifav = X amps.

Ifav is described as the maximum mean on-state current.

Since in a full rectifier bridge, every diode conducts only half the
time, would it be proper to say that I can make the bridge to rectify
2X current, out of four such diodes?

Thanks

i

 

[John Popelish]

Suppose that I build a rectifier bridge to rectify 60 Hz AC.

Each diode has a rating of Ifav = X amps.

Ifav is described as the maximum mean on-state current.

Since in a full rectifier bridge, every diode conducts only half the
time, would it be proper to say that I can make the bridge to rectify
2X current, out of four such diodes?

Thanks

i

All other things being equal (like ambient air temperature around the
diode or amount of heat sink per diode, etc.) then, yes. Each diode
in a full bridge carries half of the average output current.

 

[Ignoramus361]

All other things being equal (like ambient air temperature around the
diode or amount of heat sink per diode, etc.) then, yes. Each diode
in a full bridge carries half of the average output current.

That's great. So, to paraphrase your answer, I could build a 480 amp
rectifier bridge for balanced 60 Hz AC, out of four Ifav=240V
diodes, provided that I satisfy your conditions on equal heating etc.

Am I correct?

thanks

i

 

[Tim Williams]

That's great. So, to paraphrase your answer, I could build a 480 amp
rectifier bridge for balanced 60 Hz AC, out of four Ifav=240V
diodes, provided that I satisfy your conditions on equal heating etc.

Sure. Don't forget you'll be burning about a kilowatt in that bridge.

Wait: "balanced 60Hz AC"? As in a center-tapped winding? Cut your losses
by half and use a single pair of diodes!

Tim

 

[[email protected]]

I'm not expert on electronics, but I think what he is getting at is
that theoretically yes - but that doesn't mean you should do it.

Some things to remember, device ratings are rarely 100% accurate, there
is varience in everything. If you assume that a device is 100% precise
then you're running a risk when a real part is slightly out. Running
devices at their absolute maximum limit can reduce their life time
because you're placing more stress on them. As mentioned before -
devices typically take on different properties with temperature.... so
how it reacts are 5 celsius could be different 30 celsius. Also don't
forget that your circuit should be tollerent of outside changes beyond
your control. For example, what happens if your supply voltage surges
briefly or if the alternating frequency changes or stops alternating
(because someone connected it to an odd supply etc). You don't want
the diodes going bang as a result.

The main point is that while a lot of things can work, that doesn't
mean its advisable to cut things that close - especially when you're
dealing with power circuits.

Mike

 

[Ignoramus361]

Sure. Don't forget you'll be burning about a kilowatt in that bridge.

Thanks. Yes, it would be a lot!

Wait: "balanced 60Hz AC"? As in a center-tapped winding? Cut your losses
by half and use a single pair of diodes!

Well, by balanced I meant that 50% of time it is positive and 50% it
is negative.

Good point on the center tapped winding.

i

 

[Ignoramus361]

I'm not expert on electronics, but I think what he is getting at is
that theoretically yes - but that doesn't mean you should do it.

Some things to remember, device ratings are rarely 100% accurate, there
is varience in everything. If you assume that a device is 100% precise
then you're running a risk when a real part is slightly out. Running
devices at their absolute maximum limit can reduce their life time
because you're placing more stress on them. As mentioned before -
devices typically take on different properties with temperature.... so
how it reacts are 5 celsius could be different 30 celsius. Also don't
forget that your circuit should be tollerent of outside changes beyond
your control. For example, what happens if your supply voltage surges
briefly or if the alternating frequency changes or stops alternating
(because someone connected it to an odd supply etc). You don't want
the diodes going bang as a result.

The main point is that while a lot of things can work, that doesn't
mean its advisable to cut things that close - especially when you're
dealing with power circuits.

Sure. Real life and being careful demands some degree of derating. So
that, say, a 240 amp diode is only used for 200 amps. But, if I use
the same degree of derating, could I use diodes that could be used to
conduct 200 amps after derating, to form a full rectifier bridge to
conduct 400 amps with the same degree of safety?

i

 

[John Popelish]

That's great. So, to paraphrase your answer, I could build a 480 amp
rectifier bridge for balanced 60 Hz AC, out of four Ifav=240V
diodes, provided that I satisfy your conditions on equal heating etc.

Am I correct?

Yes. If each diode is heat sinked to meet the requirements of
Ifav=240 amps, half wave, then the bridge would be capable of 480 amps.

 

[Ignoramus361]

Yes. If each diode is heat sinked to meet the requirements of
Ifav=240 amps, half wave, then the bridge would be capable of 480 amps.

Thank you! At 1.4 voltage drop on each diode, that would mean
1.4*2*480 = 1344 watts dissipated. Quite a bit!

i

 

[John Fields]

Sure. Don't forget you'll be burning about a kilowatt in that bridge.

Wait: "balanced 60Hz AC"? As in a center-tapped winding? Cut your losses
by half and use a single pair of diodes!

 

[[email protected]]

Thank you! At 1.4 voltage drop on each diode, that would mean
1.4*2*480 = 1344 watts dissipated. Quite a bit!

And don't forget that if you are doing something evil and immoral, like
using your diode bridge to charge up a big reservoir capacitor, the
current flows through the diode for only a few milliseconds in each
cycle, so the peak current through the diode - and the peak voltage
drop - are higher than the numers you first thought of.

Diodes are relatively forgiving of peaky currents, so the bottom line
is usually the peak junction temperature of the diode under a sustained
worse case load, and answering that not only requires that you know
curent and voltage over the cycle, but also the thermal reistance to
ambient from the diode junction through any heat-sink you use to mount
the diodes to the air around the heat-sink.

People have been known to worry about the heat-capacity of the
junction.

In some semiconductors - laser diodes are the obvious example - the
junction is so small, and the heat capacity so low, that a few
microseconds of overload can heat the junction to the point of
re-annealing it into something useless for the intended purpose.

The data sheets for well-specified rectifier diodes do give a series of
absolute maximum currents for a series of progressively longer pulses,
which provides pretty much this information.

 

[Ignoramus361]

But then you need twice as much wire in the secondary (but only half
the cross sectional area) as in the case of a full bridge, and the
amount of reverse voltage the rectifiers have to stand off increases
by a factor of two.

It seems that higher power draw of the full bridge, vs. a bigger
transformer and higher reverse voltage, is a tradeoff depending of
investment cost vs. use cost, in other words depends on how much the
project item would be used in real life. Though the extra tw power
diodes would also add $60 or so to the cost.

i
--

 

[Ignoramus361]

And don't forget that if you are doing something evil and immoral, like
using your diode bridge to charge up a big reservoir capacitor, the
current flows through the diode for only a few milliseconds in each
cycle, so the peak current through the diode - and the peak voltage
drop - are higher than the numers you first thought of.

Diodes are relatively forgiving of peaky currents, so the bottom line
is usually the peak junction temperature of the diode under a sustained
worse case load, and answering that not only requires that you know
curent and voltage over the cycle, but also the thermal reistance to
ambient from the diode junction through any heat-sink you use to mount
the diodes to the air around the heat-sink.

Thanks. That makes perfect sense. For now, my interest is mainly
theoretical.

i

People have been known to worry about the heat-capacity of the
junction.

In some semiconductors - laser diodes are the obvious example - the
junction is so small, and the heat capacity so low, that a few
microseconds of overload can heat the junction to the point of
re-annealing it into something useless for the intended purpose.

The data sheets for well-specified rectifier diodes do give a series of
absolute maximum currents for a series of progressively longer pulses,
which provides pretty much this information.

Bill Sloman, Nijmegen

--

 

[Robert Latest]

On Mon, 05 Dec 2005 21:04:11 GMT,

That's great. So, to paraphrase your answer, I could build a 480 amp
rectifier bridge for balanced 60 Hz AC,

Why did I just know the OP would start pulling numbers like these out of
his hat? ;-) robert

 

[Mark]

Bill is correct...

this is a very simple circuit but very complex to analyze

If you are using a cap input filter after the bridge rectifier then the
diodes do NOT conduct 1/2 the time, they conduct much LESS then 1/2 the
time. The cap delivers the AVERAGE load current and the diodes have to
recharge the AVERAGE load current... but they only conduct for a few ms
each so the PEAK current is MUCH higer then the average and the RMS
current is also much higher than the average.

This is what power factor correction is all about.

Four diodes rated at 10 AMPs RMS used in a bridge with a cap input
filter and a small conduction angle can proably deliver considerably
less than 10 Amps DC.

This is NOT a simple question.

By the way, this low conduciton angle and high RMS current also impacts
the transformer rating...

Cap input filter power supplies are not simple to analyze.

Do some reaserch on power factor correction, conduction anlge and RMS
vs AVERAGE values.

Enjoy..

Mark

 

[Ignoramus21666]

Bill is correct...

this is a very simple circuit but very complex to analyze

If you are using a cap input filter after the bridge rectifier then the
diodes do NOT conduct 1/2 the time, they conduct much LESS then 1/2 the
time. The cap delivers the AVERAGE load current and the diodes have to
recharge the AVERAGE load current... but they only conduct for a few ms
each so the PEAK current is MUCH higer then the average and the RMS
current is also much higher than the average.

This is what power factor correction is all about.

Four diodes rated at 10 AMPs RMS used in a bridge with a cap input
filter and a small conduction angle can proably deliver considerably
less than 10 Amps DC.

This is NOT a simple question.

By the way, this low conduciton angle and high RMS current also impacts
the transformer rating...

Cap input filter power supplies are not simple to analyze.

Do some reaserch on power factor correction, conduction anlge and RMS
vs AVERAGE values.

How would one correct for it, add an inductor between the caps and
diodes?

i

Enjoy..

Mark

--

 

[Mark]

ther eis no simple answer..

theoretically, an inductor will help but such an inductor is usually
not practical for low vltage high current supplies

you can use a smaller output C which increases the ouput ripple so the
diodes conduct longer but then you need to filter or reulate the output
to get rid of the ripple..

or you can add some series R which reduces the peak current but hurts
efficency and regulation...

probably the best idea in practice is to make sure the diodes are well
over-rated


....build a breadboard and get a current probe and a scope and look at
some voltage AND CURRENT waveforms...that is the best way to get a feel
for this...maybe make a spreadsheet and take some points off the
current waveform and convert it to RMS ....try that for some simple
rectangular waveform...say a 100 Amp pulse that is 1% duty factor...
the peak current is 100 Amps,, the average current is 1 Amp....you
calculate the RMS current... the RMS current is what determines how
hot the diode will get...


Mark