Help diagnose micro rectifier failure, please

[mmagliaro]

I use an ultra tiny SMD bridge rectifier in the tender of my brass 0-6-0 project to give me a tender taillight white LED that lights in both forward and reverse.  NGineering sells them.  They are rated ate 140mA 50v - more than plenty for track voltage and LED loads.

The other day, it just flat-out failed.  The LED would light only in reverse but not forward.  I replaced the rectifier and it works again.
I presume one or more of the diodes inside the bridge burned out.

Any theories as to why this would happen?  The track power goes into the "AC" terminals of the rectifier.  The "DC" terminals of the rectifier go through a series resistor and then to the LED.    Would a rectifier this small possibly blow out from a high voltage spike from the motor during brief intermittent contact?  (Like we see LEDs get damaged when the motor field collapses during intermittent contact) ?
It is a coreless motor.  I am not sure how those behave, but there just has to be a spike of voltage coming out of the coil when power is removed.

It makes me wonder if I should have put the series resistor BEFORE the rectifier, just to current-limit any spike that hits the rectifier.

I welcome all interested theories!  Thanks!

[peteski]

Yeah, one or 2 diodes failing would cause this malfunction. If you have a multi meter with a diode test range you could actually find out whether it was just one or both diodes that died. I suspect it was only one. Sounds like you had a flawed rectifier which just failed. This can happen from time time.

Since the current limiting resistor is in series with the rest of the circuit, its location is irrelevant.  If there is a voltage spike, the elevated voltage will be seen across all the components proportionally to each components electrical properties. I suspect this was an anomaly and the replacement should have a long and happy life.

[mmagliaro]

Yeah, one or 2 diodes failing would cause this malfunction. If you have a multi meter with a diode test range you could actually find out whether it was just one or both diodes that died. I suspect it was only one. Sounds like you had a flawed rectifier which just failed. This can happen from time time.

Since the current limiting resistor is in series with the rest of the circuit, its location is irrelevant.  If there is a voltage spike, the elevated voltage will be seen across all the components proportionally to each components electrical properties. I suspect this was an anomaly and the replacement should have a long and happy life.

That's what I'm hoping.  But usually things don't work for months like this before failing.  Well, we'll know in a few more months!

[central.vermont]

Just for mere curiosity I think you should at-least contact NGineering to see if this has been mentioned to them before as a problem.

Jon

[mmagliaro]

Just for mere curiosity I think you should at-least contact NGineering to see if this has been mentioned to them before as a problem.

Jon

A reasonable suggestion.  I have sent them a note.

[Point353]

I use an ultra tiny SMD bridge rectifier in the tender of my brass 0-6-0 project to give me a tender taillight white LED that lights in both forward and reverse. 
They are rated at 140mA 50v - more than plenty for track voltage and LED loads.
The other day, it just flat-out failed.  The LED would light only in reverse but not forward.  I replaced the rectifier and it works again.
The track power goes into the "AC" terminals of the rectifier.  The "DC" terminals of the rectifier go through a series resistor and then to the LED.   
It is a coreless motor.  I am not sure how those behave, but there just has to be a spike of voltage coming out of the coil when power is removed.
It makes me wonder if I should have put the series resistor BEFORE the rectifier, just to current-limit any spike that hits the rectifier.

Rather than relocating the resistor, try adding a small surface-mount ceramic chip capacitor (on the order of 0.01-0.1uF) across the "AC" input terminals of the bridge rectifier.
If you experience another failure, there are some small surface-mount transient suppressors that could be tried.
This is one such device: http://www.littelfuse.com/~/media/electronics/datasheets/tvs_diode_arrays/littelfuse_tvs_diode_array_sd_c_datasheet.pdf.pdf

[peteski]

Make sure the ceramic chip capacitor is rated for high enough of a voltage - many of the small ones are rated  for surprisingly low voltages.

[dcutting]

I've used that bridge rectifier many times on my projects. Digikey offers two different packages of it for ~$0.40 each, just an FYI. The ones ngineering sells are marked up quite a bit. I have never had a problem with it, even when rectifying DCC and supplying a regulator that has an ATtiny841 microcontroller plus nine LED's hooked up to it for my signal decoder. I'd second the measure for a 0.1uF capacitor as well... not really necessary in my mind but it could help. Just make sure that it is rated for 25-50 volts or so. 

David

[mmagliaro]

Thanks, David.  Yes, I also have found them at Mouser now for about $0.60 each. 

Ngineering replied and said they have not had any trouble or reports of these things failing.  I described my application and circuit to them.  They don't see any problem with it. 

The cap is one of those "couldn't hurt" things, but I am reluctant to just throw components at a problem without really understanding why this would have happened in the first place (other than a one-off fluke where the rectifier was defective).

[peteski]

I would chalk this up to "infant mortality". Yes, you built the circuit months ago but how many total hours of operation did it see with voltage applied to it?  Probably just few hours.  The other possibility is that the heat stress from soldering created a microscopic fracture in the diode die or in the bond of the die to the metal which eventually failed.

[mmagliaro]

I would chalk this up to "infant mortality". Yes, you built the circuit months ago but how many total hours of operation did it see with voltage applied to it?  Probably just few hours.  The other possibility is that the heat stress from soldering created a microscopic fracture in the diode die or in the bond of the die to the metal which eventually failed.

Correct, this engine does not have a lot of run time on it.
And also correct... this is a TINY surface mount part, and it's hard to get "in and out" so you don't heat damage it.   I definitely wrecked one of them before I got it right.   So that could easily have been it.

[mmagliaro]

Sadly, it's time to resurrect this thread.  The micro rectifier has failed again.

It *has* to be an emf kick-back from the motor, I would think.  These tiny rectifiers (data sheet here: http://www.ngineering.com/CMKBR-6F.pdf  claim to have a peak reverse voltage of 60 volts.
Maybe that's not high enough?

The circuit is very simple.  The rail power goes into the AC terminals of the rectifier.  The DC outputs from rectifier
go through a resistor to an SMD LED.   And the LED never fails.  Normally, this allows the LED to light in both directions.
When the rectifier fails, it only lights in reverse, so one of the diodes in the bridge has clearly opened.

Before I go gung-ho on this and remove weight to open up space in the tender to put a physically larger rectifier in there, is there some other phenomenon that could account for this?   Could the collapsing field from a coreless motor
really put out a spike higher than 60 volts?  And if it did, wouldn't that take out the LED with it?  I've got a 1.2k resistor
in there.  So at 12v, the current is limited to 10 mA.  But at 60v, the current would soar to 50 mA and surely kill that LED.

[mmagliaro]

Sadly, it's time to resurrect this thread.  The micro rectifier has failed again.

It *has* to be an inductive kick-back from the motor, I would think.  These tiny rectifiers (data sheet here: http://www.ngineering.com/CMKBR-6F.pdf  claim to have a peak reverse voltage of 60 volts.
Maybe that's not high enough.

The circuit is very simple.  The rail power goes into the AC terminals of the rectifier.  The DC outputs from rectifier
go through a resistor to an SMD LED.   And the LED never fails.  Normally, this allows the LED to light in both directions.
When the rectifier fails, it only lights in reverse, so one of the diodes in the bridge has clearly opened.

Before I go gung-ho on this and remove weight to open up space in the tender to put a physically larger rectifier in there, is there some other phenomenon that could account for this?   Could the collapsing field from a coreless motor
really put out a spike higher than 60 volts?  And if it did, wouldn't that take out the LED with it?  I've got a 1.2k resistor
in there.  So at 12v, the current is limited to 10 mA.  But at 60v, the current would soar to 50 mA and surely kill that LED.

Perhaps the only thing that has saved the LED is that the diode in the bridge fries? 

If I put an SMD bridge in there with a 100v peak limit (or even 600v), am I also going to need a Schottky diode reverse-wired across the LED?  It seems like a logical conclusion.

[peteski]

IMO, no need for a protection diode across the LED. Not only there are 2 rectifier diodes wired in series with the LED, the other 2 of the rectifier diodes are wired in reverse parallel with the LED. Plus, the rectifier will not allow reverse voltage to even enter that circuit. What you are thinking of doing is more than a typical "belt and suspenders" solution.

400V or higher seems to also be an overkill. I would say that a motor (at full 12V applied to it, then the circuit interrupted) should not put out any more than 50V (but you seem to be experiencing more than that?).  Unfortunately it is not something you can measure without a storage oscilloscope.

I also think that the problem might be with the rectifiers themselves being marginal (weak or out of spec). There are tens of thousands of DCC decoders which drive DC motors using full voltage 12V pulses and nothing on them blows up.  I highly doubt that any of the decoder's components are rated higher than even 100V.

[mmagliaro]

Peteski,
Everything you say is pretty much what I would have thought.  Sure is weird, though.  I suppose I could have gotten a bum batch of these.

If I had a bench supply that could go up to 60v, I would take another one from the package and start driving it with higher and higher voltage to see what the failure point is.  That would be very educational.

Meanwhile.... I found one other manufacturer that makes an SOT-363 package bridge rectifier (the same tiny footprint as the one I've been using).   They have an 85 volt reverse voltage rating.  I ordered a dozen from Mouser and I'll put one of those in there when they arrive.  Using a different part from a different manufacturer with a higher voltage rating will probably be a good thing.