Why Would Flyback Secondary Diode Fail?

[nihan_A]

Hi,

I have a product that is been used in the field for years. I received a complaints from a customer who is using our products in the boats and the customer said the product does not work. When I examined the product I found that on the power supply (3W flyback converter) the secondary output diode (D50 on schematic) have been failed (not short circuit and there is no indication that shows the diode is burnt or anything. it looks fine on the surface). They returned 10 pcs of products and all of them had this problem on the power supply. When I changed this diode the product is working completely fine.

You can see the schematic of the flyback circuit below. As I stated, this product is been using in the field for years but I didn't receive any product with a problem regarding this diode. This is the first time and all of these 10pcs products is used on the boat. So I think maybe there would be something in the boats causing this issue.

Looking for any idea regarding what might cause it.

Thanks in advance.

 

[KlausST]

Hi,

* a schematic without values. Text almost impossible to read.
* not even the name of diode in question is easy to read. "SK1BA" seems not to be correct) No link to a datasheet.
* output voltage is 12V?
* no transformer information

I like to help, but can find out how.

A diode is not rated for "3W power supply output". Main spec is current(peak, average...) and reverse voltage.

A boat... sweet water or salt water?
--> Photos of PCB layout and PCB with failed diode on it.
In which application is it used? is there a possibilty of reverse voltage or externally forced current?

Klaus

 

[cupoftea]

Yes, SK3BA is only rated 100v, and it may see more than that...we wouldnt know without your ns/np and mains input.....also, BOM. Also need trr of your output diode.
Your bias coil also looks wrongly phased...link in the LNKXXX datasheet and ill chek for you.

 

[FvM]

Your bias coil also looks wrongly phased...link in the LNKXXX datasheet and ill chek for you.

It's correct, positive voltage in flyback phase.

--- Updated Aug 1, 2022 ---

Without knowing the winding ratio we can't determine, if there's even a chance to apply overvoltage to the 100V diode.

 

[cupoftea]

Oops yes sorry i misread your schem, bias is indeed correct phased.
Boats can have big input voltage transients, which may have got referred to the sec diode and blown it up.
Or maybe the snubber for it wss open cct or wrong components. Maybe RCD clamp components were wrong, and a high voltage again got referred to sec.

Your schem shows pri and sec grounded to the same gnd which is unusual.

 

[betwixt]

Grounding would be my first guess. Consider what would happen if other equipment, outside the PSU, either conducted or spiked a voltage to the output. It would be reinforced by the 'mains' (generator?) voltage at the input and almost certainly exceed the PIV rating by a large amount.

Brian.

 

[Easy peasy]

some one has changed the diode in manufacture ..... or the purcahsing people have switched to a new source but really the diode is not the same ( same root cause ).

--- Updated Aug 2, 2022 ---

Also - if the user has connected a 12V battery the wrong way round to the psy - it will kill the diode pretty quick - first going short then open - like a fuse ...

if the diode is truley open circuit - it is likely they have done this and not told you.

This seems to fit your scenario:
"When I examined the product I found that on the power supply (3W flyback converter) the secondary output diode (D50 on schematic) have been failed (not short circuit and there is no indication that shows the diode is burnt or anything. it looks fine on the surface). They returned 10 pcs of products and all of them had this problem on the power supply. When I changed this diode the product is working completely fine."

 

[dick_freebird]

I would expect reverse overvoltage to fail a diode shorted,
but overcurrent to fail it (bondwire?) open. How you'd get
gross overcurrent out of a flyback I do not know. Perhaps
full load (or over) testing? Do you have any finer information
on just what test step (or testing / application @ customer)
caused or first exposed the failure? Do you have additional
units which you could methodically abuse to replicate the
mode and its "trigger"? Then you could impose the same,
to a diode alone, and see what it takes and whether this
violates any ratings, or whether (say) you got sold some
counterfeit that doesn't live up to the imagined source's
ratings because it's reclaimed and rebranded to something
it's not.

I would recommend that you attempt to de-encapsulate
the diode and look at it under high magnification to see
if you can deduce a failure mode (this may be impractical
without a very good microscope and a wet bench). As a
part of this (and more doable w/ low gear) you'd pull I-Vs
off damaged and virgin samples. You might also get a
"known good, known source" sample de-encapsulated to
compare, and know then about the counterfeit question.

Have you proven that the failure is indeed the diode,
and not the parallel RC "snubber" elements or board
connections or secondary winding? Open circuits can be
hard to debug but chasing the secondary current loop
around its course with an ohmmeter could yield clues.

 

[Easy peasy]

the IC will not allow gross overcurrent - and even then a low V batt or low ohm load is needed for same - hence one of the 2 scenarios listed in post #2 is most likely

the small o/p diode will easily fail short then open under applied battery ... as the small junction is the weakest link in the chain.

 

[cupoftea]

Also, did you re-check the "dead" diode after you pulled it out?.....it may have gone dry joint when it was in cct.
...hence it appeared open.
Drys are , IMHO, the most common cause of problems that arent cct design errors.

Its an SMA, so youd have to push it into some blutak, so it doesnt skid away when you probe it, then be sure to probe it with non oxidised probes on the diode checker.
______________---
Also, even though you are on a boat, you will have an "earth bus" conductor....maybe someone plugged something into the flyback output which was grounded to earth.....and with your lack of isolation, that could well have killed the diode....as i see Betwixt already kindly implied above.

 

[scopeprobe]

Hi,

I have a product that is been used in the field for years. I received a complaints from a customer who is using our products in the boats and the customer said the product does not work. When I examined the product I found that on the power supply (3W flyback converter) the secondary output diode (D50 on schematic) have been failed (not short circuit and there is no indication that shows the diode is burnt or anything. it looks fine on the surface). They returned 10 pcs of products and all of them had this problem on the power supply. When I changed this diode the product is working completely fine.

You can see the schematic of the flyback circuit below. As I stated, this product is been using in the field for years but I didn't receive any product with a problem regarding this diode. This is the first time and all of these 10pcs products is used on the boat. So I think maybe there would be something in the boats causing this issue.

Looking for any idea regarding what might cause it.

Thanks in advance.


View attachment 177769

I guess theres a question regarding the Load thats attached. What is it? If it has any inductive propertites (pump, motor etc.) then a back EMF could exceed the 100V rating of the diode. Theres no significant filtering except the output diode so could possibly exceed the reverse rating of the diode.

 

[Easy peasy]

except the any back emf generated goes thru the diode in its forward direction - not the other way around - draw a diagram and figure it out ....

 

[Taihung]

You should check that on the boat, your customer will use generator or not. This mean that AC input voltage is not stable, maybe higher when they increase speed that make reverse voltage on diode higher and maybe over its maximum rating.

 

[Warpspeed]

I tend towards Easy P's theory in post 7.

Most semiconductors when over stressed, break down resulting in a shorted component.
Open circuit failures seem to be pretty rare among overloaded power semiconductors.
This sounds more like a component quality control issue rather than a circuit design problem, especially if the product has been proven reliable in the past.

These days fake and counterfeit parts are far from unknown, and a discussion with your purchasing officer and his supplier might be in order.