ESD protection circuit analysis

[Hicham M]

Hello,


I have a device that reads CAN bus information, but today the device doesn't work anymore.

I opened the device and started debugging it ( I have the schematic ).
The following schematic is the part for CAN.


CAN1_RX, CAN1_TX goes straight to a microcontroller.

I found that the dioes in D25 : D5 (between4-2) and D6(between6-2) are broken.

the PESD5V0L5UY is a Low capacitance unidirectional fivefold ElectroStatic Discharge (ESD) protection diode.

now, i'm basically a software developper, so I only have basics in electronics.

My question is : what could make these 2 diodes burned out like that ?

I admit that I messed around with the board. any advice how to avoid that next time ?

and I also noticed that the Vcl of the PESD5V0L5UY is 12V image below



I'm using a power supply of 24V, could this be the reason of my problem ??

 

[dick_freebird]

You should first tey and find the threat source or info
about it (if this is a field return).

Then, what you mean by "burned out" along with what
you know about the event, may lead you somewhere.

Those zeners / TVSes have peak current and short term
average current and short/medium term power limits. An
ESD event of classical form is very energy constrained
and is unlikely to pop a purpose-designed TVS. But there
are many ways to overstress an input which can violate
pulse energy limits and fuse a bond wire or ruin the
silicon thermally.

Much of the time an incoming "ESD failure" will be failure
analyzed and show that it's really an out-of-spec EOS
(electrical overstress) event that is responsible.

I might recommend that you place some moderate value
resistors between the input pins and the TVSes, such that
a credible worst case input overvoltage cannot drive the
TVS to a destructive operating point. Your inductors (which
show values in ohms; why?) might protect the receiver
but can't help the clamps. However moving them to outboard
of the clamps might pay off, on the cheap. Provided that
the clamp capacitance does not ruin signal qualities, when
so-limited.

 

[chuckey]

I would put small resistors on the output pins of your CANBUS. If the resistors go in series with the protection diodes, then any voltage overstress would go to the micro. I think its more likely that the CANBUS was touched on a high voltage point (+5V?). the input of the receiver would have diode clamps to ground and +3.3V hence the current limiting inductors.
Frank

 

[Hicham M]

You should first tey and find the threat source or info
about it (if this is a field return).

Then, what you mean by "burned out" along with what
you know about the event, may lead you somewhere.

Those zeners / TVSes have peak current and short term
average current and short/medium term power limits. An
ESD event of classical form is very energy constrained
and is unlikely to pop a purpose-designed TVS. But there
are many ways to overstress an input which can violate
pulse energy limits and fuse a bond wire or ruin the
silicon thermally.

Much of the time an incoming "ESD failure" will be failure
analyzed and show that it's really an out-of-spec EOS
(electrical overstress) event that is responsible.

I might recommend that you place some moderate value
resistors between the input pins and the TVSes, such that
a credible worst case input overvoltage cannot drive the
TVS to a destructive operating point. Your inductors (which
show values in ohms; why?) might protect the receiver
but can't help the clamps. However moving them to outboard
of the clamps might pay off, on the cheap. Provided that
the clamp capacitance does not ruin signal qualities, when
so-limited.

Thank you for the feedback, I mean by "burned out" broken, I think the technical term is Failed or fault... ?


About showing the inductors value in ohms, I checked the datasheet and I noticed a note : "Resistance element becomes dominantat high frequencies.", Yet this is the case for frequencies like 100MHz, while my board works at 250KHz and 500KHz, and maybe at 1MHz (max can bus frequency).
Link to the datasheet: **broken link removed**

So, u're right, mentioning the value of the inductor in ohms doesn't mean anything in my case, even the use of this inductor.

I'll check your recommandations, Thank you.

- - - Updated - - -

I would put small resistors on the output pins of your CANBUS. If the resistors go in series with the protection diodes, then any voltage overstress would go to the micro. I think its more likely that the CANBUS was touched on a high voltage point (+5V?). the input of the receiver would have diode clamps to ground and +3.3V hence the current limiting inductors.
Frank

I think you're right, I might put 24V in the CAN1H and CAN1L.
Do you think that even +5V would overstress the diodes ? it's Vcl voltage is -12V, does it mean that they support input voltage up to 12V ??